<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Arun Sahlam on Medium]]></title>
        <description><![CDATA[Stories by Arun Sahlam on Medium]]></description>
        <link>https://medium.com/@arunsahlam?source=rss-f52df2e123a5------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>Stories by Arun Sahlam on Medium</title>
            <link>https://medium.com/@arunsahlam?source=rss-f52df2e123a5------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 08:50:51 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@arunsahlam/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[When Analysis Is Cheap, Observation Is the Moat]]></title>
            <link>https://medium.com/@arunsahlam/when-analysis-is-cheap-observation-is-the-moat-12850e7dc60f?source=rss-f52df2e123a5------2</link>
            <guid isPermaLink="false">https://medium.com/p/12850e7dc60f</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[data-strategy]]></category>
            <category><![CDATA[business-strategy]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[venture-capital]]></category>
            <dc:creator><![CDATA[Arun Sahlam]]></dc:creator>
            <pubDate>Sat, 07 Nov 2026 16:41:01 GMT</pubDate>
            <atom:updated>2026-11-07T16:41:01.547Z</atom:updated>
            <content:encoded><![CDATA[<p>For the past decade, the tech playbook was clear: build a data lake, hoard proprietary records and call yourself ‘data-driven’.</p><p>When a capable model can ingest, clean and summarise most datasets quickly, owning historical records no longer guarantees an edge. What stays scarce is knowing what has not yet been recorded. The alternative is the observation moat: a defensible edge built from privileged, real-time access to how work actually happens, turned into proprietary insight faster than competitors can replicate the access.</p><h3>1. What data hoarding could and could not buy</h3><p>Two acquisitions show how heavily the industry invested in data advantage, and what that investment could and could not buy.</p><p><strong>Truven Health Analytics (2016).</strong> In February 2016, IBM announced a US$2.6 billion deal to buy Truven for its Watson Health unit. IBM said its health cloud would hold about 300 million patient lives, roughly 215 million of them from Truven. The premise was that accumulating one of the world’s largest proprietary pools of health data, paired with Watson, would create a decisive lead. In January 2022, <a href="https://www.hcinnovationgroup.com/finance-revenue-cycle/health-it-market/news/21253978/ibm-sells-off-watson-health-assets">IBM agreed to sell the Watson Health data and analytics assets</a>, Truven’s included, to Francisco Partners for an undisclosed sum (IBM had reportedly sought about US$1 billion). Truven’s core products, such as <a href="https://www.mrweb.com/drno/news33526.htm">MarketScan and Micromedex</a>, live on within the buyer’s standalone company, Merative, suggesting that the underlying data kept its commercial value. According to contemporary reporting, what fell short of expectations was the business built on the assumption that hoarded data combined with automated analysis would translate into an advantage. <a href="https://www.industryweek.com/finance/article/21970655/ibm-buys-truven-health-analytics-for-26-billion">IBM Buys Truven Health Analytics for $2.6 Billion</a></p><p>One plausible reading is that claims records capture financial billing transactions, not the friction of clinical delivery. A claims database records that a procedure occurred; it misses operational realities, such as staff who might re-enter patient notes across disconnected legacy systems or improvise around triage bottlenecks. Financial records document outcomes, but they omit the human friction that shapes them.</p><p><strong>Metaweb (2010).</strong> In July 2010, <a href="https://betanews.com/article/google-buys-metaweb-to-improve-results-for-complex-search-queries/">Google acquired Metaweb</a>, the company behind Freebase, a collaboratively edited structured database. Terms were not disclosed, but Metaweb had raised close to US$60 million, and its database catalogued some 12 million items. Metaweb does not illustrate data hoarding. It shows that structuring knowledge at scale once took a funded company and a community of contributors. Today, modern language models can extract comparable entity maps from unstructured text at very low marginal cost, although the output still needs checking.</p><p>Data moats have not vanished entirely. They persist in regulated domains, proprietary transaction networks and rare longitudinal datasets. But general data hoarding is a declining asset, with value moving from owning historical records to capturing live operational reality.</p><h3>2. The return of fieldwork</h3><p>If data analysis is cheap and abundant, what becomes rare? <strong>Understanding what has not yet been formalised as information.</strong></p><p>This economic shift is driving the return of fieldwork. The idea itself is old: lean manufacturing has long asked managers to go to the <em>gemba</em>, the real place where work happens, a practice <a href="https://www.mheducation.com/highered/mhp/product/gemba-kaizen-commonsense-approach-continuous-improvement-strategy-second-edition.html">Masaaki Imai describes in <em>Gemba Kaizen</em></a>. What has changed is the economics. When analysis is cheap, first-hand sight of the work becomes the scarcer input.</p><p>The story goes that <a href="https://en.wikipedia.org/wiki/Genchi_genbutsu">Taiichi Ohno</a> had trainees stand in a chalk circle on the factory floor and simply watch, though even that tale is credited to him only ‘perhaps apocryphally’.</p><p>The most valuable insights do not live in tidy executive dashboards; they hide in shadow workflows:</p><ul><li>The custom spreadsheet nobody admits is mission-critical.</li><li>The manual reconciliation process executed every Friday afternoon.</li><li>The late-night phone call placed to resolve an emergency exception.</li><li>The sticky note taped to a monitor, recording an ad hoc workaround for software that never quite fitted the actual job.</li></ul><p>An observation moat compounds through three drivers:</p><ol><li><strong>Access density:</strong> Each workplace where a team earns continuous, trusted access reveals new operational exceptions to learn from.</li><li><strong>Tacit knowledge:</strong> Each exception teaches the team what subtle indicators to look for next. That know-how builds up inside the team and rarely reaches a document a competitor could copy.</li><li><strong>Relationship depth:</strong> Front-line operators will show a trusted partner the workarounds they would hide from an enterprise software vendor.</li></ol><p>Nonaka and Takeuchi’s The Knowledge-Creating Company draws the classic line between tacit knowledge, which people hold but rarely write down, and explicit knowledge, which can be recorded.</p><p>Traditional analytics dashboards leave all of this contextual reality out by design.</p><h3>3. Seeing the whole chain</h3><p>Much enterprise software measures results, which are lagging indicators recorded long after the work has concluded. A different approach starts earlier in the operational sequence:</p><p>Context → Action → Result</p><p>Technologies like digital twins and human–AI teaming can help by making process drift and hidden bottlenecks visible. A digital twin lets a team compare how work is supposed to run with how it actually runs, and that gap is where the next observation begins. Rather than reviewing a lagging output after a failure, these systems can turn ground-level signals into a living picture of how operations run, and may expose structural gaps before dashboards register a drop in performance.</p><h3>4. Applying the shift: logistics and exception handling</h3><p>Consider complex operational environments like warehouse fulfilment and logistics exception handling.</p><p>Viewed purely through a traditional data lens, operations are evaluated through aggregate throughput metrics: items scanned per hour, average transit times and automated sorting rates. These metrics optimise for what can be easily counted, but they can mask systemic friction on the floor.</p><p>In an observation-led model, AI handles the structural data load: tracking asset locations, parsing shipping manifests and modelling baseline supply flows. People observe what spreadsheets miss. Picture a dock crew that routinely bypasses one automated scanner, damaged packaging dealt with out of camera shot, or a layout constraint that forces a manual workaround. None of this would appear in a throughput report, and each is a clue to where the system design and the real work have drifted apart. Organisations that pair AI-driven analytics with direct observation of these exceptions are harder to copy than those relying on dashboard telemetry alone.</p><p>Natan Linder and Trond Arne Undheim make a related case for human-centred frontline operations in <a href="https://www.wiley.com/en-be/shop/general-introductory-business-management/augmented-lean-a-human-centric-framework-for-managing-frontline-operations-p-9781119906018"><em>Augmented Lean</em></a>. Linder is the CEO of Tulip, a frontline operations software company.</p><h3>5. Defending the moat</h3><p>A sceptical reader will naturally ask: <em>If observation is simply watching workflows, can’t any competitor do the exact same thing?</em></p><p>Recording an observation turns it into data, but the moat is not the record. It is the access that produced it and the judgement that read it, neither of which a competitor can download. Anyone can casually tour a factory floor or a distribution centre. Staying close to the people who handle operational exceptions, week after week, is harder. So is having the domain expertise to recognise which small anomaly signals a larger shift before anyone else does. That combination is what makes the moat defensible.</p><p>Scaling beyond a single observer is the harder question. One route is to build observation into the software itself and to staff a team whose job is to audit where the real workflow breaks away from the system design, rather than only to analyse outputs.</p><p>Palantir offers one example. Its forward-deployed engineers sit inside customers’ organisations, often for months, to learn the domain and build workflows on the spot, as <a href="https://www.a16z.com/the-palantirization-of-everything">a16z describes</a>. The model is costly, and a16z notes that the talent and culture behind it don’t generalise easily. Embedded observation can compound, but it doesn’t scale cheaply.</p><p>Observation has its own failure modes: operators may alter their behaviour when watched, observers are constrained by where they gain entry, and confirmation bias can distort field interpretations. For this reason, observation and data must work together. Observation without data is anecdote; data without observation is abstraction. The enduring competitive advantage belongs to those who use privileged ground-level access to interpret the data before the wider market recognises what to look for.</p><h3>Putting it together</h3><p>Builders can start before they build. Spend time inside the workflow before writing code, instrument the exceptions rather than the happy path, and treat each spreadsheet and sticky note as a specification.</p><p>If a founder claims an observation moat, three questions help.</p><ol><li>Who lets you watch, and why would they keep letting you?</li><li>What have you learned that no dataset you could buy contains?</li><li>How does that learning reach the product without depending on one person?</li></ol><h3>The mandate for builders</h3><p>When information is abundant and analysis is cheap, understanding becomes rare.</p><p>The task for the next decade is to step away from the aggregate dashboard, stay close to the real work and build systems grounded in operational truth. Pay attention to the sticky notes on the monitor.</p><p><em>What is the sticky note in your organisation? I’d like to hear about it in the comments.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=12850e7dc60f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[We Say We Teach History for Its Patterns. Mostly, We Don’t.]]></title>
            <link>https://medium.com/@arunsahlam/we-say-we-teach-history-for-its-patterns-mostly-we-dont-74e77f266cbd?source=rss-f52df2e123a5------2</link>
            <guid isPermaLink="false">https://medium.com/p/74e77f266cbd</guid>
            <category><![CDATA[teaching]]></category>
            <category><![CDATA[history]]></category>
            <category><![CDATA[critical-thinking]]></category>
            <category><![CDATA[education]]></category>
            <category><![CDATA[psychology]]></category>
            <dc:creator><![CDATA[Arun Sahlam]]></dc:creator>
            <pubDate>Sat, 31 Oct 2026 16:56:01 GMT</pubDate>
            <atom:updated>2026-10-31T16:56:01.595Z</atom:updated>
            <content:encoded><![CDATA[<p><em>History’s patterns are probabilistic, its causes contested and its mechanisms psychological. What would teaching it that way look like?</em></p><p>Ask most people why schools teach history and you will get a shrug and a platitude: ‘learning from the past’, ‘knowing where we came from’. The better answer is that history is not a stock of facts to be recited. It is a working model of how causes connect to outcomes, revised as evidence and questions change. The facts are the raw material; the explanatory work lies in the connections between them.</p><p>That answer survives the obvious objection — everything is online, so why memorise dates? — but it comes with an uncomfortable corollary. History as it is usually taught is dominated by dates, actors and chronology, and the reasoning that supposedly justifies the subject is what gets squeezed out. Where schools do attempt causal reasoning, they seem to omit the layer that would make it rigorous: psychology. Fixing this means facing three problems in turn: the ideal itself, the institutional incentives that undermine it, and the psychological layer we leave out.</p><h3>The Index Is Not the Education</h3><p>Knowing that the Treaty of Versailles was signed in 1919 is an index lookup, not an education. A student who can trace how its terms fed German grievance, how that grievance was exploited, how institutional weakness and economic collapse made extremism viable, and how contingent choices at several forks produced a catastrophe has learnt something transferable.</p><p>Search can tell you what happened. These days an AI model will also offer you a why, fluently and with confidence. What neither will do, unless you already know to ask, is tell you how sure we should be, or which of several competing explanations holds up best against the evidence. The value lies in the model, not the data points.</p><p>None of this makes facts irrelevant. Working memory needs anchor facts to recognise patterns in real time, and without a baseline chronology a student cannot formulate a counterfactual or spot an anomaly. Facts are the beginning of historical thinking, not the end of it.</p><p>A reference work is static by design, which is what makes it useful. History as a discipline is dynamic: new evidence, new questions and new methods continually reweight the model. Teaching it as a finished product misrepresents it.</p><h3>Patterns, Not Physics</h3><p>Historical regularities are real, but they are not laws of nature. ‘Economic crisis breeds extremism’ is a useful heuristic, not a prediction rule. It is context-dependent, mediated by institutions, contingent on human choice and routinely defeated by cases that do not fit.</p><p>Teach historical patterns as if they were physics equations and you get the worst outcome: a confident determinism that mistakes a tendency for a guarantee. The patterns are probabilistic and defeasible, and their value lies in sharpening judgement, not licensing prophecy. A student who leaves history class believing they can predict the next collapse from a checklist has been miseducated, even if every item on the checklist is accurate.</p><p>This is why the probabilistic wording matters. ‘This made an outcome more likely’ is a different and more honest claim than ‘this caused the outcome’, and learning to ask how much more likely, and on what evidence, is a genuine intellectual skill.</p><h3>Causation Is Overdetermined and Contested</h3><p>Most significant historical outcomes have several sufficient causes, and serious historians legitimately disagree about how to weight them. Was the First World War chiefly the product of alliance systems, imperial rivalry, nationalist mobilisation or a series of miscalculations? Was the 1947 Partition of India the result of colonial divide-and-rule policies, institutional failures, or political miscalculations by nationalist elites? Each answer captures something real; none exhausts the event.</p><p>So the skill students need is not to learn the causal chain that led to X, but to construct and critique causal arguments: to hold several explanations in view at once, to ask what evidence would discriminate between them, and to recognise when an explanation is doing ideological work rather than explanatory work. That is a more demanding target than the tidy version, and it is the actual point.</p><p>It is also a civic necessity. A citizen who does not know how institutions, wars and economic crises unfolded cannot evaluate claims about them. A citizen who knows only one sanitised causal story is worse off still — fluent in the narrative, defenceless against it. Democracies depend on people who can recognise propaganda, spot recurring patterns and interrogate the story they are being told, including the one the curriculum itself offers.</p><h3>The Institutional Friction: Why Classrooms Fall Short</h3><p>Now the uncomfortable part. The case above is largely aspirational. In many school systems — from Western curricula to India’s exam-driven Board system — history teaching tends to be dominated by names, dates, key terms and a chronological march through events — the very things the justification says matter least. That holds even where, as in England, the National Curriculum names ’cause and consequence’ among the concepts pupils must understand and <a href="https://www.gov.uk/government/publications/research-review-series-history">Ofsted’s 2021 history research review</a> treats causation as a core disciplinary concept. The reasons seem mostly structural rather than intellectual.</p><p><strong>Assessment drives much of it.</strong> Causal and probabilistic reasoning is hard to mark reliably and at scale. ‘Was the primary cause of the First World War the alliance system or imperial rivalry?’ produces answers that are expensive to evaluate and invite dispute. ‘In what year was the Treaty of Versailles signed?’ is a clean, defensible, machine-markable item. So the testable fragments crowd out the untestable whole — an effect assessment researchers call ‘washback’. History teachers in England have described it: writing in <em>Teaching History</em>, <a href="https://www.history.org.uk/publications/resource/8132/assessment-after-levels">Geraint Brown and Sally Burnham</a> warned that using GCSE mark schemes lower down the school simply encourages teaching to the test, because those schemes largely ignore substantive knowledge and the complexity of second-order conceptual development.</p><p><strong>Chronology is the cheap prerequisite.</strong> You cannot reason about causal relationships without knowing what happened and in what order, so the scaffolding gets mistaken for the building. Teachers under time pressure teach the prerequisite and never reach the payoff.</p><p><strong>Textbooks are built for coverage.</strong> James Loewen’s survey of 12 US history textbooks, <a href="https://thenewpress.org/books/lies-my-teacher-told-me/"><em>Lies My Teacher Told Me</em></a> (1995), argued that they supply masses of detail while omitting pivotal questions. The same logic plausibly applies elsewhere. A textbook must touch every mandated topic, so it compresses causation into a sentence or two per event and spends its space on names, dates and key terms in bold. The relational structure, which is the real intellectual content, is sacrificed for space.</p><p><strong>Curricula swing between knowledge and skills.</strong> One camp says students must know the facts; another says they must think like historians. The relational approach sits awkwardly in the middle, because it needs real content and sophisticated reasoning. Each camp tends to under-serve it: content advocates may distrust it as soft, and skills advocates may lack the content depth to teach it well.</p><p><strong>Teachers are often trained the same way.</strong> Many history teachers were themselves taught the subject as a sequence of events, and reproducing that is the path of least resistance as research on teacher socialisation suggests. Teaching causal structure well takes a teacher who has internalised historiographical debate, which is a high bar.</p><p>Taken together, these pressures suggest the gap is not an oversight. It is what tends to happen when a discipline’s actual method is expensive to teach and to assess, and a system optimises for coverage and defensible marks. The talk of ‘critical thinking’ and ‘learning from the past’ may then be a justification applied after the fact to a practice organised around the transmission of an easily tested chronological narrative.</p><p>None of this means nobody is trying. There is a rich body of research on ‘historical thinking’ — including work by Sam Wineburg and colleagues at the <a href="https://inquirygroup.org/">Digital Inquiry Group</a> (formerly the Stanford History Education Group) — and the US Advanced Placement history courses name causation, comparison and contingency among their reasoning skills. The argument here is that the pressures above work against such efforts, and that psychology is seldom made explicit even in the better ones — a claim worth testing against actual curricula.</p><h3>What Teaching History as Causal Reasoning Looks Like</h3><p>If the framing were taken seriously, the classroom would move away from ‘what happened’ towards five habits:</p><ul><li><strong>Counterfactuals</strong>, not as parlour games but as a discipline for isolating which causes were load-bearing. If X had not happened, how robustly does the outcome change?</li><li><strong>Competing explanations</strong>: assigning two historians who disagree and asking students to adjudicate on evidence, not authority.</li><li><strong>Probability language</strong>: teaching students to say ‘this made X more likely’ rather than ‘this caused X’, and to ask how much more likely, and on what basis.</li><li><strong>Mechanism tracing</strong>: the intermediate steps between a condition and an outcome, which coverage-driven teaching skips.</li><li><strong>Cross-context comparison</strong>: did similar conditions elsewhere produce similar outcomes? Some courses assess comparison — AP World History treats it as one of three reasoning skills — but comparison used as an explicit check on a causal claim, rather than a list of similarities and differences, is rarer.</li></ul><p>All five require students to reason about how an outcome was produced, not merely that it was. And every one runs into the same wall: the mechanisms are psychological.</p><h3>The Missing Layer: Why History Needs Social Psychology</h3><p>When a historian writes ‘grievance fed extremism’ or ‘humiliation radicalised the public’, they are making psychological claims without the vocabulary or evidence base to defend them. The causal links in much political history bottom out in how people perceive threats, conform to groups, rationalise the status quo, defer to authority, and respond to scarcity or humiliation. Without that layer, ‘cause’ is a black box labelled with a verb.</p><p>Social psychology is a mechanism library for much of political history. Concepts such as Henri Tajfel’s in-group and out-group dynamics, Stanley Milgram’s obedience to authority (whose methodological flaws and protocol deviations were detailed in Gina Perry’s 2012 study, <a href="https://thenewpress.org/books/9781595589217/">Behind the Shock Machine</a>), Ziva Kunda’s motivated reasoning and Albert Bandura’s moral disengagement supply the micro-foundations that historical causal claims silently depend on.</p><p>Bringing the two together would do four things.</p><p><strong>It would discipline counterfactuals.</strong> When students ask ‘would things have gone differently if X?’, psychology constrains the answer by showing which human variation is plausible and which is not. A causal story that requires people to behave in psychologically implausible ways is a bad causal story.</p><p><strong>It would help explain recurrence.</strong> Why do similar patterns — scapegoating, panic, mass mobilisation — appear across wildly different societies and centuries? A plausible answer is that the underlying cognitive and social machinery is stable, and that stability is what makes historical patterns probabilistic rather than arbitrary. On this view the patterns recur because the people are, in relevant respects, the same.</p><p><strong>It would correct the ‘great man’ default.</strong> A psychological lens directs attention to crowd behaviour, cognitive bias, elite panic and the decisions of ordinary people, which is where much of history actually happens.</p><p><strong>It would make the probability honest.</strong> ‘This made extremism more likely’ is a claim about human response tendencies under particular conditions, and it cannot be taught honestly without the psychology that generates it. The two are not separate subjects that happen to be adjacent: the psychological layer is the mechanism the probabilistic language points at.</p><p>The interwar period shows the difference. A typical textbook compression runs: ‘Germany faced economic ruin, which led to the rise of the Nazi Party and eventually the Second World War.’ That sentence is accurate in outline and almost entirely uninformative.</p><p>A psychologically informed version might read: ‘Status loss and perceived humiliation made many citizens available to extremist mobilisation, and institutions under stress proved unable to resist it.’ It is a hypothesis rather than a settled finding, but it names the machinery. It tells you what to look for elsewhere, and it can be tested against other cases — including the evidence on who actually voted for the Nazis, which historians dispute. Richard Hamilton’s <a href="https://press.princeton.edu/books/hardcover/9780691642017/who-voted-for-hitler"><em>Who Voted for Hitler?</em></a> (1982), for instance, found support concentrated among the better-off, which sits awkwardly with a status-loss story, while Thomas Childers’s<a href="https://uncpress.org/9780807841471/the-nazi-voter/"> <em>The Nazi Voter</em></a> (1983) argued for a broader, catch-all constituency. A hypothesis that meets evidence like this has to be refined or abandoned, and that is the method this essay is arguing for.</p><p>Perpetration is a different mechanism from mobilisation. Christopher Browning’s <a href="http://harpercollins.com/products/ordinary-men-christopher-r-browning"><em>Ordinary Men</em></a> (1992), a study of a German reserve police battalion that took part in mass shootings of Jews in occupied Poland, is often read as showing how peer pressure, deference to authority and moral disengagement can turn unremarkable men into perpetrators.</p><p>Daniel Goldhagen’s <a href="https://www.penguinrandomhouse.com/books/68841/hitlers-willing-executioners-by-daniel-jonah-goldhagen/"><em>Hitler’s Willing Executioners</em></a> (1996) offered a very different account, attributing such killing to a distinctively German ‘eliminationist antisemitism’ — an ideological explanation rather than a situational one. Set side by side, the two books let students ask what evidence would discriminate between the explanations. Treated this way, history becomes evidence bearing on a mechanism rather than a story with a moral.</p><h3>Avoiding the Traps: Psychohistory and Social Mechanisms</h3><p>This integration has failure modes that must be named, or it will deservedly lose credibility.</p><p><strong>Presentism and anachronism.</strong> Applying modern clinical categories to past actors is a well-known error. ‘Was X a psychopath?’ is usually a category error dressed as analysis. The lens has to be about general mechanisms, not retrospective diagnosis of individuals.</p><p><strong>Psychohistory’s bad reputation.</strong> The tradition of psychoanalysing dead leaders from fragmentary evidence earned its disrepute. The credible version relies on well-supported social-cognitive findings rather than speculative depth psychology.</p><p><strong>The field’s own replication problems.</strong> Psychology has faced well-publicised replication difficulties: in the <a href="https://doi.org/10.1126/science.aac4716">Open Science Collaboration’s 2015 project</a>, about 36 per cent of 100 replication attempts reached statistical significance, against 97 per cent of the original studies. Classic studies have also been heavily criticised —<a href="https://doi.org/10.1037/amp0000401"> Thibault Le Texier’s archival work (2019)</a> argues that the guards in the Stanford Prison Experiment were told how to behave, and aspects of early obedience research have been challenged too. Any teaching that borrows from psychology must say which findings are robust, which are contested and why, applying to psychology the same critical standards brought to history.</p><p><strong>Over-psychologising structural causes.</strong> Psychology explains mechanisms, not the whole causal story. Institutions, economies and contingencies still do heavy lifting. The risk is a reductionism that says ‘it was all human nature’, which is unfalsifiable and politically convenient, since it lets structures off the hook.</p><p><strong>Confirmation bias in mechanism-matching.</strong> Once you have a toolkit of biases, you can find one to explain anything. The discipline has to be to ask whether the evidence actually shows this mechanism operating here, or whether the match is being made after the fact. Psychological explanation is seductive precisely because it is so flexible, and a theory that can explain anything is of little scientific use.</p><p>Beyond these theoretical traps lies a practical obstacle: the academic silo. Departments, training, journals and timetables are separate, and career incentives reward depth in one field rather than fluency across two. A history teacher who wants to teach motivated reasoning needs enough psychology to do it without caricature; a psychology teacher who wants to use historical cases needs enough history to avoid treating them as convenient illustrations. Current training may produce neither by default.</p><h3>Teaching the Machinery, Not the Narrative</h3><p>We claim to teach history for its causal patterns and critical thinking, but in many classrooms we do not, and the reasons look institutional rather than intellectual. Even where the causal version is attempted, the psychological mechanisms that would make it rigorous tend to be left out, so the causal claims float free of the evidence that would support them.</p><p>That is a less flattering thesis than ‘history teaches us to learn from the past’, but a truer one, and it changes the curriculum question. It is no longer ‘which facts must students know?’ but ‘which patterns are worth modelling, which psychological mechanisms generate them, and how do we teach students to test those mechanisms rather than assume them?’</p><p>We don’t teach history so students can recite what happened. We teach it so they can reason about why, and tell the difference between an explanation and a story someone wants them to believe. That takes the causal reasoning we claim to value, the psychology we currently omit, and the institutional courage to assess something more expensive than a date. Once we make that shift, the curriculum becomes a live historical question in its own right, open to the same evidence, revision and dispute as the history it teaches.</p><p><em>How did your own history lessons treat causes: as a chain to memorise, or an argument to test? Teachers’ and students’ experiences would be especially welcome in the responses.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=74e77f266cbd" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Plans Fail: Ten Dimensions Method for Finding the Real Cause]]></title>
            <link>https://medium.com/@arunsahlam/why-plans-fail-ten-dimensions-method-for-finding-the-real-cause-57f611f702c9?source=rss-f52df2e123a5------2</link>
            <guid isPermaLink="false">https://medium.com/p/57f611f702c9</guid>
            <category><![CDATA[self-improvement]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[problem-solving]]></category>
            <category><![CDATA[project-management]]></category>
            <dc:creator><![CDATA[Arun Sahlam]]></dc:creator>
            <pubDate>Sun, 25 Oct 2026 04:41:01 GMT</pubDate>
            <atom:updated>2026-10-25T04:41:01.704Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Score ten dimensions, map how they interact, and fix the cause rather than the symptom.</em></p><p>Picture this. You start a fitness programme with a detailed plan, a committed mindset and money for the gym. Three weeks later, you’ve quietly stopped going. Or you lead a product launch with a talented team, thorough market research and adequate funding, and watch it underperform badly.</p><p>In both cases, the useful question isn’t just ‘what went wrong?’ It’s ‘why did reality diverge from our expectations?’ Without a structured answer, we tend to repeat the same mistakes, apply surface-level fixes such as ‘try harder’ or ‘communicate more’, or lose confidence in our ability to plan at all.</p><p>This article introduces a ten-dimension framework for examining that gap. It sorts the possible causes into individual, operational and contextual factors, then walks you through a five-step process for gathering evidence, scoring each dimension, mapping how the dimensions interact and choosing targeted interventions. It works for a stalled career move as well as a struggling project, and for teams as well as individuals.</p><h3>What You’ll Be Able to Do</h3><p>By the end of this article, you will be able to:</p><ul><li>Apply the ten dimensions to a real deviation from your professional or personal life</li><li>Tell individual, operational and contextual causes apart</li><li>Spot where dimensions compound, offset or hide one another</li><li>Turn your findings into targeted interventions and a reusable checklist</li></ul><p>No specialist background is needed. If you have ever asked ‘why didn’t that work?’ after a project, goal or initiative, you can use this framework. It helps to have one real example in mind to work through as you read.</p><h3>The Framework at a Glance</h3><p>Here, a <em>deviation</em> is simply the gap between the outcome you expected and the outcome you got. The ten dimensions are ten different lenses for explaining that gap.</p><p><strong>Individual factors</strong></p><ul><li><strong>1. Motivation:</strong> Did people want to succeed, and were they rewarded for it? <em>Warning sign: effort fades after the initial push.</em></li><li><strong>2. Capability:</strong> Did they have the skills and knowledge required? <em>Warning sign: repeated errors, slow progress.</em></li><li><strong>3. Intention:</strong> Were the goals clear, shared and genuinely wanted? <em>Warning sign: people pulling in different directions.</em></li><li><strong>4. Readiness:</strong> Were they prepared, and was the timing right? <em>Warning sign: rushed start, missing prerequisites.</em></li></ul><p><strong>Operational factors</strong></p><ul><li><strong>5. Resources:</strong> Did they have the materials, funding, tools and support? <em>Warning sign: constant workarounds, stalled purchases.</em></li><li><strong>6. Capacity:</strong> Was there enough time and bandwidth to do the work? <em>Warning sign: overtime, dropped tasks.</em></li><li><strong>7. Reliability:</strong> Did the systems and support structures hold up? <em>Warning sign: recurring breakdowns, last-minute failures.</em></li><li><strong>8. Process/Procedure:</strong> Was the method suitable, and was it actually followed? <em>Warning sign: confusion over who does what, inconsistent results.</em></li></ul><p><strong>Contextual factors</strong></p><ul><li><strong>9. Integrity:</strong> Were commitments genuine and results reported honestly? <em>Warning sign: surprises that surface late, sanitised reporting.</em></li><li><strong>10. Environment:</strong> Did external conditions help or hinder? <em>Warning sign: an outside shock derails the plan.</em></li></ul><p>The dimensions overlap at the edges, which is why each one below includes a short note on how to tell it apart from its neighbours. Where two seem to fit, record the evidence under the one that explains it best, and capture the link between them in Step 4 of the process.</p><p><strong>A quick test for the family:</strong> if the fix is a change in the person (skills, wants, preparation), the cause is Individual. If the fix is a change in structure or supply (budget, schedule, tools, workflow), it is Operational. If the cause lies in the ethical climate of the effort, or outside its boundary, it is Contextual.</p><h3>Individual Factors</h3><p>The human element is the foundation of any outcome, whether personal or professional. These four dimensions ask whether people wanted to, could, aimed at the right target, and were prepared at the right moment. A quick way to remember them: <strong>want</strong> (Motivation), <strong>can</strong> (Capability), <strong>aim</strong> (Intention), <strong>ready</strong> (Readiness).</p><h3>1. Motivation: Did people have the drive and the incentives to succeed?</h3><p>Motivation covers both what people feel and what their surroundings reward. A sales team might miss its targets not because of market conditions, but because bonuses were tied to short-term metrics that discouraged the slower work of building relationships. On a personal level, someone might drop a language-learning goal because pressure from family overshadowed any genuine interest in the culture.</p><p><strong>Key questions:</strong> What was rewarded, formally and informally? Did personal values align with the goal? Was there intrinsic motivation beyond external pressure?</p><p><strong>Tell it apart:</strong> Ask whether effort would return if the rewards and pressures changed. If so, it is Motivation. If effort would still lag because people disagreed about, or did not truly want, the goal itself, it is Intention (dimension 3).</p><h3>2. Capability: Did they have the skills and knowledge required?</h3><p>A marketing manager with deep experience in traditional advertising may struggle to run digital campaigns. Someone changing careers may lack the specific technical skills the target role demands. High motivation cannot close a fundamental capability gap; it only means people struggle harder.</p><p><strong>Key questions:</strong> Were skills assessed beforehand? Did team members receive the training they needed? Which competencies were assumed but never verified?</p><p><strong>Tell it apart:</strong> Capability asks whether people could do the task today. Readiness (dimension 4) asks whether the preconditions for the plan were in place.</p><h3>3. Intention: Were the goals clear, shared and genuinely wanted?</h3><p>Misaligned intentions often masquerade as execution problems. An organisation may announce a market expansion while key stakeholders privately prefer stability. An individual may chase a promotion because they feel they should, while privately valuing the free evenings the new role would take away.</p><p><strong>Key questions:</strong> Did everyone share the same definition of success? Were conflicting priorities acknowledged openly? Was the ‘why’ compelling enough to survive the first obstacle?</p><p><strong>Tell it apart:</strong> Ask whether fully motivated people would all be pulling in the same direction. If not, it is an Intention problem. Integrity (dimension 9) is different: it asks whether people were honest in how they pursued and reported on the goal.</p><h3>4. Readiness: Were they prepared, and was the timing right?</h3><p>Readiness is distinct from capability (what you can do) and intention (what you want to do). It asks whether preparation was complete and the moment was right. A sales team may have the skills and the drive but still not be ready if it hasn’t yet absorbed a new pricing model. A runner with a sound plan isn’t ready to start it while recovering from illness.</p><p><strong>Key questions:</strong> Were the prerequisites in place? Was the timing appropriate? Had the necessary preparations, such as training, onboarding or handovers, been completed?</p><p><strong>Tell it apart:</strong> Readiness is about preconditions and timing, not skill. If people could not have done the task even with perfect preparation, it is a Capability problem.</p><h3>Operational Factors</h3><p>Even willing, skilled and aligned people can fail inside structures that don’t support them. These four dimensions look at what surrounds the person: what they had, how much they could handle, whether it held up, and how the work was organised.</p><h3>5. Resources: Were sufficient materials, funding, tools and support available?</h3><p>Resources are what you have to work with. They include tangible assets and intangible support such as mentorship, sponsorship or organisational backing. An individual might abandon a fitness goal because gym fees are unaffordable, while a project team might struggle with inadequate software licences or sudden budget cuts.</p><p><strong>Tell it apart:</strong> If the fix is ‘we needed more of something’, it’s a Resources problem.</p><h3>6. Capacity: Was there enough bandwidth to do the work effectively?</h3><p>Capacity is about throughput: the time, energy and attention available to use what you have. A person might have the money for a course but no mental bandwidth during a stressful period. A department might have an adequate budget but too little operational capacity during peak season or while several initiatives run at once.</p><p><strong>Key questions:</strong> What else was competing for time and attention? Were workloads realistic? Was there slack for the unexpected?</p><p><strong>Tell it apart:</strong> If the problem is ‘we had the means but not the room to use them’, it’s Capacity, not Resources.</p><h3>7. Reliability: Did the systems and support structures hold up?</h3><p>Reliability is about consistency over time. Personally, it might mean steady daily routines or dependable childcare. Organisationally, it covers dependable technology, predictable team performance and stable processes. A fitness plan can fail not for lack of motivation, but because the bus to the gym is never on time.</p><p><strong>Key questions:</strong> What broke, stalled or arrived late? How often did it happen? Were there backups?</p><p><strong>Tell it apart:</strong> Resources asks whether something existed. Reliability asks whether it worked every time you needed it.</p><h3>8. Process/Procedure: Was the method appropriate, and was it actually followed?</h3><p>This dimension asks two separate questions: was the method suited to the task, and did people really use it? Poor personal productivity systems can undermine capable individuals, and unclear workflows can cripple talented teams. Good processes let ordinary performers deliver consistently; poor ones waste the talent of even the best.</p><p><strong>Key questions:</strong> Was there a defined method? Was it fit for purpose? Where did people depart from it, and why? Were handoffs and decision rights clear?</p><p><strong>Tell it apart:</strong> Process asks whether the method was suitable and followed. If the method was sound but nobody had time to follow it, that is a Capacity problem.</p><h3>Contextual Factors</h3><p>The final two dimensions step back to look at the ethical climate in which the effort took place and the external conditions surrounding it.</p><h3>9. Integrity: Were commitments genuine and results reported honestly?</h3><p>Integrity is about honest conduct and honest reporting: whether commitments were real, shortcuts were avoided and results were described truthfully. For individuals, integrity problems show up as shortcuts that undermine long-term success, or as telling yourself the plan is on track while ignoring the evidence. In organisations, they create hidden problems: unreported conflicts of interest, data massaged to meet a deadline, or promises to stakeholders that can’t be kept.</p><p>Integrity failures deserve special attention because they are concealed by nature. They tend to surface late, and they make the evidence for every other dimension less trustworthy.</p><p>To detect an integrity problem, triangulate self-report against independent sources: compare status updates with logs, timestamps and financial records; check whether bad news travelled upward quickly; and ask whether people felt safe saying that something was off.</p><p><strong>Key questions:</strong> Were status reports accurate? Were any commitments made that people knew couldn’t be kept? Were corners cut, and did anyone feel unable to say so?</p><p><strong>Tell it apart:</strong> Integrity concerns honesty in conduct and reporting. A goal that nobody truly wanted is an Intention problem, not an Integrity one.</p><h3>10. Environment: Did external conditions help or hinder?</h3><p>Environmental factors sit largely beyond your control but profoundly affect outcomes. Personal examples include family circumstances, economic conditions and life events. Professional ones include market changes, regulatory shifts and organisational culture. The same plan can succeed in one environment and fail in another.</p><p>Because you can rarely change the environment, the intervention here is usually to adapt: adjust timing, narrow the scope or build in contingency. (Some elements, such as organisational culture, can be influenced over time.)</p><p><strong>Key questions:</strong> What changed outside our control while the plan was running? Were those changes foreseeable? Did we have a contingency?</p><p><strong>Tell it apart:</strong> Environment covers what changed outside the effort’s control. A risk that everyone could see in advance but nobody acted on belongs to Readiness or Process.</p><p><strong>Watch out:</strong> Environment is the easiest dimension to blame precisely because it is outside your control. Check that you have honestly examined the other nine before settling on it.</p><h3>Applying the Framework: A Five-Step Process</h3><p>Effective application requires systematic evaluation rather than intuition. Follow these steps. They are not strictly linear: Steps 2 and 3 in particular are iterative, so expect to return to evidence-gathering as unknowns surface.</p><h3>Step 1: Define the expected outcome</h3><p>Write down what success looked like at the outset: the specific metrics, the timeframe and the assumptions behind them. Then record the actual outcome and the gap between the two. Without this baseline, ‘failure’ is just a feeling.</p><h3>Step 2: Gather evidence for each dimension</h3><p>Collect objective evidence through data, documents, surveys, interviews and observation. Prefer what happened over what people remember or assume. For personal goals, calendars, activity logs, messages and bank statements are better witnesses than memory. Be wary of hindsight bias: once you know the outcome, the causes always look more obvious than they were.</p><p>Not all evidence is equal. In rough order of strength: hard data (logs, timestamps, financial records), direct observation, structured interviews, and finally self-report or memory. Back every important claim with at least two kinds, and treat a claim that rests on self-report alone as provisional.</p><h3>Step 3: Score each dimension</h3><p>Rate each dimension from 1 to 5, where 1 is a major weakness that clearly contributed to the gap, 3 is mixed or unclear, and 5 is a strength that was not a factor. Write one line of evidence beside every score. If you cannot cite evidence, mark the dimension ‘unknown’ rather than guessing; an unknown is a signal to return to Step 2. Where possible, have two people score independently and then discuss any differences.</p><p>Beside each score, note when the dimension degraded: from the start, at a particular moment, or gradually. ‘Never ready’ and ‘drifted over time’ call for different interventions. Also watch for the actor–observer asymmetry, a well-known if debated tendency: people often explain their own failures by circumstances (Environment) and other people’s failures by traits (Motivation, Capability). Having someone outside the effort score independently helps counter it.</p><h3>Step 4: Map the interactions</h3><p>Dimensions rarely fail alone. For each weak score, ask how it connects to the others. Four patterns appear again and again:</p><ul><li><strong>Compounding:</strong> two weaknesses reinforce each other in a loop. Overloaded staff (Capacity) make more errors, and the resulting rework adds to the overload. Break the loop at its shared root.</li><li><strong>Cascade:</strong> one failure triggers the next, in sequence. A capacity squeeze leads to missed sessions, and missed sessions erode motivation. Interrupt the chain at its earliest link.</li><li><strong>Compensating:</strong> a strength covers for a weakness, at least for a while. High motivation can offset a moderate capability gap, strong environmental support can offset limited resources, and excellent process can absorb reliability problems.</li><li><strong>Masking:</strong> one problem hides another. Optimistic status reporting (Integrity) can make a broken Process look healthy until it is too late.</li></ul><p>A simple list of arrows, such as ‘capacity squeeze → missed sessions → eroded motivation’, is enough to make the chain visible.</p><p>Step 4 can also send you back to Step 3. If mapping reveals that two dimensions were really one, or that a low score was a symptom rather than a cause, revise the scores before moving on.</p><h3>Step 5: Prioritise interventions</h3><p>Don’t simply fix the lowest score. Separate <em>gating</em> dimensions, where the plan cannot succeed if they fail (no budget, a missing skill the task requires, stakeholders pulling in opposite directions), from <em>contributing</em> dimensions, which raise or lower the result without deciding it. Fix gating dimensions first: a gating dimension scoring 2 usually outranks a contributing one scoring 1.</p><p>Within each group, focus on the dimensions with the lowest scores and the greatest influence on the gap. Where one weakness drives others, fix that root first. Foundational dimensions such as Capability, Resources and Intention usually come before downstream symptoms such as a dip in motivation. For each priority, define the intervention, an owner, a measure of success and a review date.</p><p>Here is what targeted interventions can look like for each dimension, alongside the warning sign that points to it.</p><ul><li><strong>Motivation</strong> (effort fades after the initial push): realign incentives with the behaviour you want; connect the task to personal values.</li><li><strong>Capability</strong> (repeated errors, slow progress): assess skills, provide training, pair with an experienced colleague or coach.</li><li><strong>Intention</strong> (people pulling in different directions): hold an alignment session; agree a one-sentence definition of success.</li><li><strong>Readiness</strong> (rushed start, missing prerequisites): delay the start, or set explicit entry criteria before beginning.</li><li><strong>Resources</strong> (constant workarounds, stalled purchases): reallocate budget, secure tools or sponsorship, or reduce scope.</li><li><strong>Capacity</strong> (overtime, dropped tasks): limit work in progress; remove or postpone competing commitments.</li><li><strong>Reliability</strong> (recurring breakdowns, last-minute failures): add backups, maintenance routines or redundancy.</li><li><strong>Process/Procedure</strong> (confusion over who does what, inconsistent results): simplify and document the workflow; clarify decision rights.</li><li><strong>Integrity</strong> (surprises that surface late, sanitised reporting): create safe reporting channels; verify the data; reset commitments honestly.</li><li><strong>Environment</strong> (an outside shock derailed the plan): build contingency; stage decisions in gates; revisit assumptions regularly.</li></ul><h3>A Worked Example (Illustrative)</h3><p>Consider the fitness programme from the opening. A first-time gym-goer commits to five sessions a week, starting in the busiest month of the working year. By week three, they have stopped. Here is how the five steps play out.</p><h3>Steps 1 and 2: Define the gap and gather evidence</h3><p>The expected outcome was five sessions a week for 12 weeks. The actual outcome was eight sessions in three weeks, then none. To find out why, the analyst pulled four sources. The workout log showed which sessions were missed, and on which days. The work calendar showed that the first two weeks were packed with deadlines and that no gym sessions had ever been blocked out.</p><p>The bus timetable, set beside the arrival times jotted in the log, showed that a trip scheduled to take 30 minutes had often taken an hour. And a one-week time diary showed little free time left after work and travel: on the heaviest days, the gym slot came after a 12-hour day.</p><p>Reliability was initially marked ‘unknown’: the log recorded missed sessions but not why. The timetable and the time diary filled the gap. That is the loop from Step 3 back to Step 2 in action.</p><h3>Step 3: Score independently, then reconcile</h3><p>Two people scored the dimensions independently: the gym-goer and a friend acting as accountability partner. They disagreed on one dimension, Readiness. The friend scored it 3, because the membership and kit were in place. The gym-goer scored it 2. They settled on 2, because the calendar showed no sessions scheduled and the start date fell in a month already known to be the busiest of the year. The person could do the exercises, which is Capability, but the preconditions for the plan had not been assembled, which is Readiness.</p><p>The reconciled scorecard follows, with a note on when each dimension degraded.</p><ul><li><strong>Motivation: 4.</strong> Keen at the start; the goal was self-chosen and personally meaningful. <em>When it degraded:</em> held up in weeks 1–2; dipped in week 3 after missed sessions.</li><li><strong>Capability: 4.</strong> Knew the exercises and could do them safely. <em>When it degraded:</em> stable throughout.</li><li><strong>Intention: 4.</strong> One clear goal, with no competing aim. <em>When it degraded:</em> stable throughout.</li><li><strong>Readiness: 2.</strong> Started in a month already known to be the busiest of the year, with no sessions pencilled into the calendar. <em>When it degraded:</em> from day one (never ready).</li><li><strong>Resources: 4.</strong> Membership already paid for. <em>When it degraded:</em> stable throughout.</li><li><strong>Capacity: 1.</strong> Five sessions plus 60-hour work weeks left no time or energy. <em>When it degraded:</em> from week 1; worst in week 2.</li><li><strong>Reliability: 2.</strong> The bus was unpredictable, turning a one-hour round trip into two on several gym days. <em>When it degraded:</em> intermittent from week 1.</li><li><strong>Process/Procedure: 2.</strong> The plan jumped from no exercise to five sessions a week with no build-up, and had no fallback when a session was missed. <em>When it degraded:</em> broke down in week 2, at the first missed session.</li><li><strong>Integrity: 4.</strong> Workouts were logged honestly, so the log showed the decline plainly. <em>When it degraded:</em> stable throughout.</li><li><strong>Environment: 3.</strong> A client escalation extended the work crunch by two weeks beyond anything planned. <em>When it degraded:</em> hit in week 2.</li></ul><h3>Step 4 in practice: the interactions</h3><p>The instinctive diagnosis would be ‘not enough discipline’, which is a Motivation story. But Motivation scored 4 and was the last dimension to degrade. The map shows two patterns. Capacity and Reliability compounded: a packed schedule and an unpredictable bus each made every session costlier. Then a cascade began: with no fallback process, the first missed session became a missed week, and motivation eroded as a result rather than as a cause. The shared root is Readiness: the plan started at full intensity in a known-busy month, with no cushion. Capacity is the gating dimension, because no amount of motivation produces sessions that the schedule cannot hold.</p><p>Readiness and Process sit close together here, so the analyst applied the ‘record the evidence under the one that explains it best’ rule. Timing and missing preconditions (a known-busy month, nothing scheduled) went under Readiness, because starting later would have fixed them. The shape of the plan itself (no build-up, no fallback) went under Process, because only redesigning the plan would fix it.</p><h3>Step 5 in practice: the priorities</h3><p>Capacity and Readiness come first, because they gate the plan. Process comes next, to interrupt the cascade at its earliest link. Reliability comes last, because it contributes to the outcome without deciding it. The temporal column sharpens each choice. Readiness was ‘never ready’, so the fix is to delay the start, not to repair drift. Process broke down at a specific moment, so the fix is a fallback for exactly that moment.</p><ol><li><strong>Capacity and Readiness.</strong> Cut to three 30-minute sessions a week and start the Monday after the work crunch ends. <em>Owner:</em> the gym-goer. <em>Measure:</em> three sessions completed in each of four consecutive weeks. <em>Review:</em> end of week 4.</li><li><strong>Process.</strong> Add a 20-minute home routine as the fallback for any missed session. <em>Owner:</em> the gym-goer, with the friend as accountability partner. <em>Measure:</em> no week with zero sessions. <em>Review:</em> weekly check-in.</li><li><strong>Reliability.</strong> Switch to a gym within a ten-minute walk of home or work. <em>Owner:</em> the gym-goer. <em>Measure:</em> round trip under 30 minutes. <em>Review:</em> end of week 2.</li></ol><h3>Common Pitfalls</h3><ul><li><strong>Blaming individuals first.</strong> The individual dimensions are the most visible, so they attract blame by default. Examine the operational and contextual dimensions before concluding that people simply didn’t try.</li><li><strong>Treating the dimensions as independent.</strong> The most valuable insights come from the connections between them, so don’t skip Step 4.</li><li><strong>Hindsight bias.</strong> Anchor every score in evidence gathered at the time, not in a story constructed after the outcome.</li><li><strong>Reaching for Environment.</strong> It is the convenient explanation because it absolves everyone. Use it only after the others have been tested.</li><li><strong>Over-engineering minor misses.</strong> For small or one-off deviations, a quick pass through the ten core questions is enough. Reserve full scoring for deviations that are costly or keep recurring.</li></ul><h3>Using the Framework Before the Fact</h3><p>The same ten lenses work as a forward-looking check. Before launching a plan, imagine it has already failed and ask which dimension is the most likely culprit. Then strengthen that dimension now, while it is cheap to do so.</p><p>The framework also works on positive deviations. When a result beats expectations, running the ten dimensions can show you what actually made the difference, so you can repeat it deliberately rather than by luck.</p><h3>Applying the Framework to Itself</h3><p>This method can fail like any other plan. Its Motivation may be low, because few people feel like running a ten-dimension analysis after a setback. Its Capacity may be the binding constraint: you may have 20 minutes, not a day. Its Integrity risk is scoring to confirm the story you already believe. Its Environment risk is that it assumes you can gather evidence and score honestly, which may not hold in a low-trust organisation where data is unreliable or people fear speaking up. Match the depth to the situation: use the full process for deviations that are costly or recurring, and the quick pass through the ten core questions in ‘The Framework at a Glance’ when time is short.</p><h3>Build Your Own Checklist</h3><p>To turn the framework into a reusable tool for your own context:</p><ol><li>Copy the ten dimensions from ‘The Framework at a Glance’ into a spreadsheet or notes app, and add fields for Score, Evidence and When it degraded.</li><li>Add one or two questions specific to your field under each dimension. A software team might ask ‘Were dependencies on other teams confirmed?’ under Reliability; someone changing careers might ask ‘Have I spoken to people who do this job?’ under Readiness.</li><li>Add fields for Intervention, Owner and Review date for any dimension scoring 2 or below.</li><li>Revisit the checklist after each significant deviation and prune the questions that never surface anything useful.</li></ol><h3>Final Thoughts</h3><p>Plans often fail for more than one reason. Several smaller weaknesses line up, frequently in dimensions nobody was looking at. A structured, multi-dimensional review replaces the quick, comfortable explanation with an evidence-led one, and that makes the response to a setback more precise.</p><p>The value of the method lies less in the scoring itself than in what it prompts you to do: fix root causes instead of symptoms. It can help individuals set more realistic goals, and organisations form more accurate expectations and choose better remedies. Used regularly, in planning cycles as well as post-mortems, it can turn each setback into information that improves the next plan.</p><p>The next time a plan goes astray, resist the first explanation that comes to mind. Apply the ten dimensions, gather the evidence, and fix the cause rather than the symptom.</p><h3>Further Reading</h3><p>The framework sits alongside several established models. These are worth exploring if you want to go deeper:</p><ul><li><a href="https://sites.gsu.edu/7150-hpt-manual/eight-models/gilberts-behavioral-engineering-model/">Thomas Gilbert, <em>Human Competence: Engineering Worthy Performance</em> (1978), whose Behavior Engineering Model</a> separates a person’s own repertoire (knowledge, capacity, motives) from the environment around them (information, resources, incentives). It is widely used to argue that environmental gaps account for more performance problems than individual ones, which supports the ‘blaming individuals first’ warning above.</li><li><a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC3096582/">Susan Michie, Maartje van Stralen and Robert West, ‘The behaviour change wheel: A new method for characterising and designing behaviour change interventions’</a> (<em>Implementation Science</em>, 2011), which sets out the COM-B model of capability, opportunity and motivation.</li><li><a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC1117770/">James Reason, <em>Human Error</em></a> (1990), which set out the theory of active and latent failures, later popularised as the ‘Swiss cheese’ model. It is the classic account of how several small weaknesses line up into a failure, which is this article’s central thesis.</li><li><a href="https://link.amazon/B01XrYrBE">Daniel Kahneman, <em>Thinking, Fast and Slow</em></a> (2011), on hindsight bias and related distortions in how we judge past outcomes.</li><li><a href="https://www.researchgate.net/publication/313610424_The_actor_and_observer_Divergent_perceptions_of_the_causes_of_behavior">Edward Jones and Richard Nisbett, ‘The Actor and the Observer: Divergent Perceptions of the Causes of Behavior’</a> (1971), the classic account of the actor–observer asymmetry.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=57f611f702c9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Hierarchy of Thought]]></title>
            <link>https://medium.com/@arunsahlam/the-hierarchy-of-thought-77e4987f084e?source=rss-f52df2e123a5------2</link>
            <guid isPermaLink="false">https://medium.com/p/77e4987f084e</guid>
            <dc:creator><![CDATA[Arun Sahlam]]></dc:creator>
            <pubDate>Thu, 22 Oct 2026 06:32:25 GMT</pubDate>
            <atom:updated>2026-10-22T06:32:25.216Z</atom:updated>
            <content:encoded><![CDATA[<p>Spotting a pattern is not the same as using it</p><p>Imagine a chess player with six months of serious study beating someone who has played casually for five years. Or a founder with no MBA outmanoeuvring seasoned executives. Stories like these are usually filed under talent or luck. They are better explained by what each person does with what they know, and by how well that knowledge has been tested against reality.</p><p>The framework below rests on a distinction that everyday language blurs: spotting a pattern is not the same as using it. It has four levels, and it is written for professionals and leaders who want to find their own bottleneck and practise their way past it. We will follow one hypothetical person through one problem, then return to the puzzle above.</p><h3>The hierarchy at a glance</h3><ol><li><strong>Pattern.</strong> A regularity in the world: something that repeats in time, in space, or in cause and effect.</li><li><strong>Knowledge.</strong> Patterns that have been noticed, stored and organised so they can be found again when needed.</li><li><strong>Applied cognition.</strong> The ability to identify which patterns are at work in the situation in front of you, whether by recalling a known one or discovering a new one.</li><li><strong>Strategic thinking.</strong> The ability to select and combine patterns to reach a goal.</li></ol><p>Each level depends on the one below it. You cannot recall what you never stored. You cannot diagnose a situation without something to diagnose it against. And you cannot choose a sound strategy for a situation you have misread. Everyday speech lumps levels three and four together as ‘intelligence’ or ‘being smart’. Separating them is useful because people are often far stronger at one than the other: the analyst who diagnoses brilliantly but cannot decide what to do, or the decisive manager who acts with confidence on a misdiagnosis.</p><p>Problem-solving is the practical act of running the whole stack on a real problem. Proficiency is what we call the result when it runs well: reliable, adaptive, high-quality performance in a domain, achieved through calibrated pattern application. It is not a fifth rung. It is the emergent quality of the stack working, overseen by a capacity we will meet shortly: the ability to judge when a pattern can be trusted.</p><h3>A problem to follow</h3><p>Here is a hypothetical case. Priya runs customer support at a mid-sized software company. Every few weeks the ticket backlog balloons, response times collapse, and customers complain. Twice now her answer has been to hire. The backlog shrinks, then balloons again. We will watch her think about this at each level.</p><h3>Level 1: Pattern</h3><p>A pattern is any regularly repeated arrangement, sequence or set of relationships. Three kinds matter here. <em>Temporal</em> patterns repeat over time: seasonal demand, or the lifecycle of a software release. <em>Spatial</em> patterns concern arrangement in space: a city grid, the structure of a molecule, the layout of a chessboard. <em>Relational</em> patterns capture how things affect one another: if this, then that; more of one thing, less of another. Relational patterns tend to be the hardest to see, because causes rarely announce themselves the way shapes and cycles do.</p><p>Different domains lean on different kinds. Chess is largely spatial and relational: where the pieces stand and what that makes possible. Medicine is largely relational and temporal: which symptom leads to which, and how quickly. Music and programming are largely temporal and spatial: sequence and structure. People also have lopsided strengths. You may spot cycles in your workflow with ease and still consistently miss the cause and effect between colleagues.</p><p>Patterns are the building blocks of human understanding, but they are not a guarantee of truth. Three traps are common. <em>Apophenia</em> is seeing patterns in randomness. <em>Spurious correlation</em> mistakes two things that move together for two things that influence each other. <em>Overfitting</em> is learning the quirks of a small sample as if they were laws. And not every pattern is natural: some are institutional habits or inherited power arrangements wearing the costume of ‘how things are done’. Part of the skill is asking whether a regularity is real or merely familiar.</p><p>Priya’s first pattern is temporal, and she nearly misses it. She has been watching total ticket volume. A colleague suggests plotting the backlog spikes against the calendar, and they line up with product releases.</p><h3>Level 2: Knowledge</h3><p>Noticing a pattern, storing it and retrieving it are three different acts, and the weak link differs from person to person. <em>Recognition</em> is noticing a regularity in the moment. <em>Encoding</em> is storing it in a form that lasts. <em>Retrieval</em> is finding the right pattern when a new situation calls for it. Knowledge is the library that encoding builds and retrieval draws on.</p><p>Two refinements make the library metaphor more accurate. First, experts do not so much store more facts as store larger units. Classic studies of chess players by <a href="http://matt.colorado.edu/teaching/highcog/fall8/cs73.pdf">Chase and Simon</a> found that masters remember real game positions far better than novices do, yet show a much smaller advantage, or none, with random arrangements of pieces. The masters have compressed meaningful configurations into single ‘chunks’, which leaves attention free for other work. Second, there are two kinds of knowing. <em>Declarative</em> knowledge is knowing that: the rule, the named pattern, the explicit description. <em>Procedural</em> knowledge is knowing how: the executable skill that no longer needs conscious thought. The long road from knowledge to proficiency is largely the conversion of the first into the second. Knowledge that is never converted stays what <a href="https://plato.stanford.edu/entries/whitehead/">Alfred North Whitehead</a> called inert: possessed, but not used.</p><p>A library also decays without use, and old patterns can interfere with new ones. Maintenance and pruning are part of knowing, not afterthoughts.</p><p>This is where the familiar types of thinker belong. A domain expert holds a deep collection of patterns in one niche; a generalist holds a broad collection across many. ‘Common sense’ is the shared baseline library: patterns so widely held that we forget we ever learned them, which is also why it fails when we step outside the context that produced it. Neither depth nor breadth wins by default.</p><p>Priya has years of support experience, so she has seen release-driven spikes before. What she has not done is retrieve them. She filed them under ‘busy periods’, a label that points to no cause and suggests no action.</p><h3>Level 3: Applied cognition</h3><p>Applied cognition is diagnosis: deciding which patterns are at work here. A doctor recognising a cluster of symptoms is doing it. So is a salesperson reading a customer’s hesitation, or an engineer who sees that a bug belongs to a known class.</p><p>It begins with attention. Recognition is not passive reception; patterns do not announce themselves. Two people can look at the same chessboard or the same dashboard, and one sees relationships the other literally cannot perceive, because expertise is largely a matter of knowing where to look: which features are diagnostic and which are noise. This is also why experts often struggle to explain what they see. Their attention was tuned by experience, not by instruction.</p><p>Diagnosis is also where thinking most often fails quietly. A confident diagnosis of the wrong problem is worse than an honest ‘I’m not sure’.</p><p>Priya’s breakthrough is a reframing. Her problem is not too few agents. It is that every release creates a burst of confusion her team must absorb. She states the diagnosis in a form that could be proved wrong: if it is true, the tickets in each spike should cluster around the features that changed. She checks, and they mostly do.</p><h3>Level 4: Strategic thinking</h3><p>Strategic thinking is applying patterns: selecting and combining them to achieve a goal. The doctor who recognised the illness is now choosing a treatment plan, the combination of therapies, timing and follow-up that fits this patient. The salesperson who read the customer is now crafting the pitch.</p><p>Its most sophisticated form works on meta-patterns, the patterns that exist between other patterns. A technologist who sees not just one bug but the structural weakness that produces a whole class of bugs is working at this level. Meta-patterns are also the main <a href="https://www.sciencedirect.com/science/article/abs/pii/S1475158513000696">carriers of transfer</a>, the bringing of a pattern from one domain into another. Near transfer, between similar contexts, is common. Far transfer, between dissimilar ones, is rare and hard, which is exactly why it is so valuable when it happens.</p><p>Priya combines three patterns: the temporal one (spikes fade within days of a release), the relational one (changed features generate confusion tickets), and a capacity one (permanent headcount is a poor fit for temporary surges). She generates three options before choosing: keep hiring; work with the product team on clearer release notes and in-app guidance; or pre-build answer templates and a temporary release-week rota. She chooses a mix of the second and third. She also spots a meta-pattern: any change that alters user behaviour, whether a redesign, a pricing change or a policy update, creates temporary demand on the people who must explain it. She will now see it coming in places that have nothing to do with software releases.</p><h3>Calibration and metacognition: knowing when to trust a pattern</h3><p>Nothing so far guarantees that the patterns are right. What does is feedback, in the form of a prediction-and-check loop. A pattern is only as good as its record of forecasting what happens next. Without an error signal, people accumulate false patterns and never correct them; someone can spend years ‘learning’ and mostly reinforce their mistakes.<a href="https://pubmed.ncbi.nlm.nih.gov/31598236/"> Ericsson’s research on deliberate practice</a> makes a compatible point: what builds expertise is not hours but structured practice with feedback. This framework is complementary, because it asks what that practice is building and how you would know it is working.</p><p>Above the four levels sits metacognition: knowing when to trust a pattern and when to override it. This is why proficiency is not simply speed. Pattern recognition makes experts fast, but a surgeon pausing before an irreversible cut, or a strategist refusing to pattern-match prematurely on an unfamiliar situation, is being deliberately slow. Both are signs of expertise.</p><p>Consider the seasoned executive who fails during a market disruption. She lacks neither knowledge nor diagnostic skill. She fails because she applies a pattern that has stopped holding, and because her confidence has outrun her calibration. Her library is vast but uncalibrated, and nothing in her habits flags that its conditions have changed.</p><p>Before the next release, Priya writes down a prediction: with clearer notes and templates, the spike will be visibly smaller, and if it is not, her diagnosis is wrong or incomplete. Two weeks later the result is mixed. Tickets about the redesigned settings page fall sharply. Tickets about a billing change do not. The pattern was real but partial: billing confusion has a different cause that she had not seen. She does not conclude that she was wrong. She updates the library, and that update is what calibration looks like.</p><h3>Where the hierarchy holds, and where it strains</h3><p><strong>Domain validity.</strong> The framework works best where feedback is clear and quick and the environment is regular enough to learn from: chess, surgery, firefighting. <a href="https://eric.ed.gov/?id=EJ859788">Kahneman and Klein’s work on the conditions for intuitive expertise</a> draws the same line. In low-validity domains such as stock-picking or long-range political forecasting, the apparent patterns may be mostly noise, and a larger library of them produces confident incompetence rather than expertise.</p><p><strong>Breadth and depth.</strong> Specialists tend to dominate stable, high-validity domains. Generalists may have the edge in novel, ill-structured ones, where carrying a relational pattern in from elsewhere can be the advantage.</p><p><strong>Order of learning.</strong> The hierarchy describes the structure of competence, not a universal sequence of acquisition. In procedural domains such as sport, people often learn to act before they can articulate the patterns. In creative work, breaking patterns is as much a skill as seeing them.</p><p><strong>Groups.</strong> Teams and organisations have libraries too, in the form of playbooks and procedures, and their own failures of recognition, such as groupthink and shared blind spots. A team’s strategic thinking is only as good as the diversity of its attention, and a group’s library can entrench as easily as it can inform.</p><p><strong>Fuel.</strong> The framework is cognitive, but practice is not. Motivation and identity decide whether anyone sustains the habits described below, and confidence calibration decides whether strategic thinking runs without arrogance or paralysis. The hierarchy is necessary, not sufficient.</p><h3>Back to the puzzle</h3><p>The opening scenarios were illustrations, not laws. In most domains, experience usually wins, because accumulated patterns matter. The cases are interesting because they are exceptions, and the framework suggests why they happen.</p><p>The novice may have a smaller library, but each pattern in it has been tested, and a good teacher has pointed their attention at the features that matter. The five-year player may have years of untested casual play: habits rather than calibrated patterns, some of them entrenched enough to block better solutions. Experiments on the <a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC4033246/">Einstellung effect</a>, including studies with chess players, show experienced players fixating on a familiar solution and missing a better one.</p><p>The founder probably carried relational patterns over from another field and attended to features the executives had learned to ignore. Startups are also a low-validity domain, so some of the time the founder was simply lucky, and the story rarely says so.</p><p>The lesson is not that knowledge matters less than application. It is that knowledge matters in proportion to how well it has been calibrated, and that a large library is only an advantage when it is paired with the skill and the supervision to use it.</p><h3>Putting it to work</h3><p><strong>Starting from zero.</strong> A true beginner cannot yet see the patterns, so the first step is not independent discovery but borrowing someone else’s attention. Apprenticeship, mentors, named patterns from textbooks and worked case studies all do this. Learn the names first, then learn to recognise them, then begin to discover your own.</p><p><strong>Practices with a built-in check.</strong></p><ul><li><em>Keep a prediction journal.</em> Write down one pattern you see, the outcome you predict from it, and the condition under which it would fail. Review it in two weeks and score yourself.</li><li><em>Before each major decision,</em> write the pattern you are relying on and the condition under which it would stop holding.</li><li><em>Generate three approaches</em> before settling on one.</li><li><em>Hunt for meta-patterns.</em> Each month, ask which of your recurring problems are really the same problem.</li><li><em>Refresh and prune.</em> Revisit old patterns on a schedule and ask whether they still hold.</li></ul><p><strong>Locate your bottleneck.</strong> Self-report is unreliable, so these statements are behavioural. If you cannot honour one without hedging, that is probably your level.</p><ul><li><em>Pattern:</em> I can name three recurring patterns in my weekly work without consulting my notes.</li><li><em>Knowledge:</em> When I meet a new problem, I can name the specific past situation I am comparing it to, and where that comparison breaks down.</li><li><em>Applied cognition:</em> Before acting, I can state what kind of problem I face and what evidence would show I have misdiagnosed it.</li><li><em>Strategic thinking:</em> For each major decision last month, I can name the patterns I combined and the condition under which the plan would have failed.</li></ul><p>Also check for blind spots by pattern type. Are you strong on temporal patterns and weak on relational ones, or the reverse?</p><p><strong>Metrics worth keeping.</strong> Two heuristics are worth tracking, and neither comes with a universal threshold: the hit rate of your written predictions, and the share of your decisions in which you named the failure condition in advance. Compare yourself with your own baseline, not with anyone else’s.</p><h3>One last exercise</h3><p>Take the decision you are currently avoiding. Name the pattern you think governs it. Name the condition under which that pattern would fail. Now decide.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=77e4987f084e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[From Student to Candidate: 6 Habits That Make You Hireable]]></title>
            <link>https://medium.com/@arunsahlam/from-student-to-candidate-6-habits-that-make-you-hireable-72d716e05b12?source=rss-f52df2e123a5------2</link>
            <guid isPermaLink="false">https://medium.com/p/72d716e05b12</guid>
            <dc:creator><![CDATA[Arun Sahlam]]></dc:creator>
            <pubDate>Mon, 19 Oct 2026 09:40:31 GMT</pubDate>
            <atom:updated>2026-10-19T09:40:31.373Z</atom:updated>
            <content:encoded><![CDATA[<p><em>A simple system for building visible, verifiable proof of your skills, one habit at a time.</em></p><p>Picture two graduates with the same degree, the same grades and the same GitHub account.</p><p>In the interview, both are asked: <em>“Tell me about a time something went wrong and how you handled it.”</em></p><p>One freezes and says, “Hmm, I’m not sure.” The other describes a specific bug, what she tried, who she asked for feedback and what she changed afterwards.</p><p>The difference isn’t talent. It’s that the second graduate kept a record. Employers hire for <strong>communication, problem-solving and adaptability</strong>, and you can build evidence of all three long before your first interview.</p><p>This guide gives you a system for doing that. You don’t need to do everything at once. Pick <strong>one habit</strong>, run it for two weeks, then add a second. Consistency beats perfection.</p><blockquote><em>💡 </em><strong><em>A note on scope:</em></strong><em> The examples here (GitHub, pull requests, LeetCode) lean towards tech and data students. If you’re in business, humanities or marketing, swap the tools and keep the principles. Use a Notion portfolio or personal website instead of GitHub. Use case-study analyses instead of LeetCode. Use pro-bono consulting, published writing or volunteering for a society instead of open-source pull requests.</em></blockquote><h3>1. What Employers Actually Want</h3><p>A 2026 GMAC survey of 600+ recruiters worldwide ranked <strong>communication (64%), problem-solving (62%) and adaptability (60%)</strong> as the most valued skills <strong>[</strong><a href="https://www.gmac.com/market-intelligence-and-research/market-research/corporate-recruiters-survey/the-corporate-recruiters-survey-2026-report/the-corporate-recruiters-survey-2026-dashboard?dr=20"><strong>link</strong></a><strong>]</strong>. “Using AI tools” came 14th (50%). Employers assume you can prompt an AI. They want to know whether you can <em>check its output</em>.</p><p>The World Economic Forum makes a similar point <strong>[</strong><a href="https://reports.weforum.org/docs/WEF_Artificial_Intelligence_and_the_Future_of_Entry_Level_Work_2026.pdf"><strong>link</strong></a><strong>]</strong>: 37% of young workers globally are in occupations with medium-to-high exposure to AI-driven task change, and entry-level roles in the most AI-exposed quartile show nearly three times the net skills change of non-entry-level roles; employers are shifting demand toward critical thinking, problem-solving, and effective collaboration with AI agents.</p><blockquote><strong><em>Remember this:</em></strong><em> Your degree proves you can learn. Employers want to see what you’re </em>still<em> learning, what you’ve built, and what you can prove.</em></blockquote><h3>2. Six Habits to Build</h3><h4>Habit 1: Reflect every week</h4><p><strong>Start here:</strong> this Sunday, write three lines.</p><p>Every Sunday evening, open a notes app and answer:</p><ol><li>What did I do this week? (Concrete actions, not “studied.”)</li><li>What problem did I hit, or what feedback did I get?</li><li>What will I change next week?</li></ol><p><strong>Why it works:</strong> “Tell me about a time you overcame a difficulty” is a near-certain interview question, and most students freeze because they never recorded their difficulties. After three months, your log is a ready-made bank of interview stories. It also trains you to treat feedback as information, not judgement.</p><p><strong>Pass mark:</strong> four consecutive weeks. At week four, reread week one. If you can’t see any change, your entries were too vague.</p><h4>Habit 2: Step into real scenarios</h4><p><strong>Start here:</strong> book one informational interview this month.</p><p>Watching tutorials is easy and useful, but it isn’t enough. You also need live situations with real feedback.</p><ol><li><strong>Join one live project or hackathon per semester.</strong> Open-source issue threads and volunteer tech support for a small charity count too. Employers care that you’ve solved a problem with unclear conditions. Listen first, then propose.</li><li><strong>Do one informational interview per semester</strong> with someone in a non-technical role, such as a product manager, an operations lead or a salesperson. Ask how they make decisions and how they see technical colleagues.</li><li><strong>Pick learning resources by goal, not whim.</strong> Choose one channel per direction and finish a course before switching:</li></ol><ul><li>General programming: freeCodeCamp, MIT OpenCourseWare, Traversy Media, The Net Ninja</li><li>Data, AI and ML: StatQuest, 3Blue1Brown, Alex The Analyst</li><li>DevOps and cloud: TechWorld with Nana, NetworkChuck</li><li>Cybersecurity: The Cyber Mentor, John Hammond, Professor Messer</li></ul><p><strong>Close the loop:</strong> one channel, one course, one project, one portfolio entry, one LinkedIn post.</p><h4>Habit 3: Add feedback to your learning</h4><p><strong>Start here:</strong> send your next piece of writing to someone more experienced.</p><p>Most students curate and create, but skip checking and correcting.</p><ol><li><strong>Give every output a verification source.</strong> Send your article or answer to someone more senior and ask, “What’s the biggest flaw here?” The learning happens in the revision.</li><li><strong>Tag what you learn with where you’d use it.</strong> Don’t just write “learned X.” Write the scenario where X applies and what you’d do instead if you didn’t use it. This is how employers want skills described: in scenarios, not lists.</li><li><strong>Time-box your write-ups.</strong> Give yourself two hours, at most, to turn what you learned into a piece of content. The limit forces a digested summary instead of copied notes.</li></ol><h4>Habit 4: Move from posting to owning</h4><p><strong>Start here:</strong> answer one question in a community you already use.</p><p>Answering questions and making open-source contributions are good starts. The next step is taking responsibility for something small.</p><ol><li><strong>Raise your responsibility every six months.</strong> Go from answering a question, to maintaining an FAQ, to reviewing others’ pull requests, to being the contact for a small module. Maintainers remember contributors they’ve worked with, and that familiarity can help when you apply.</li><li><strong>Count non-code work.</strong> Translation, documentation, testing and event organisation are all visible contributions. Projects such as WordPress list them explicitly in their contributor pathways.</li><li><strong>Use the right platform for the right job.</strong> Use Stack Overflow, GitHub Discussions, Dev.to and Discord/Slack communities for Q&amp;A. Keep LeetCode, HackerRank and Codeforces on a separate practice track. Reddit and Quora are good for learning but weak as a showcase.</li></ol><h4>Habit 5: Treat certifications as entry tickets</h4><p><strong>Start here:</strong> write down which project your next certification would support.</p><ol><li><strong>Prefer scenario-based assessments.</strong> Microsoft Applied Skills uses lab scenarios, and IBM SkillsBuild is project-oriented. Ask: after this, will I have something I can show someone?</li><li><strong>One per semester is plenty, and it must link to a project.</strong> The order is project first, then spot the skill gap, then certify, then return to the project.</li><li><strong>Watch for the “free course” trap.</strong> Microsoft Learn courses are free, but the exam usually isn’t. Some exams, such as Oracle’s OCI AI Foundations Associate, have been offered free. Always confirm current availability in your region before you plan around it.</li><li><strong>Free places to start:</strong> Microsoft Learn Student Hub, Google Cloud Skills Boost, AWS Skill Builder, IBM SkillsBuild.</li></ol><h4>Habit 6: Be readable in three minutes</h4><p><strong>Start here:</strong> put one project README online.</p><ol><li><strong>Build a public résumé as your base.</strong> Use GitHub Pages or a clean public Notion page. Aim to finish it by the end of your second year.</li><li><strong>Make every project README answer three questions.</strong> What problem does it solve (not just which technology it used)? How do I run it (one command or clear steps)? What did I learn, and what was the hardest bug?</li><li><strong>Treat LinkedIn as a thinking showcase, not a résumé backup.</strong> Every two weeks, post what you learned from a project or certification and what you’ll do differently next time.</li><li><strong>Prepare a three-minute self-pitch</strong> covering what you’re doing now, what problem you solved, and what you want next.</li><li><strong>Optional extras, used with care.</strong> A Google Scholar profile only makes sense once you have a paper or preprint. An Upwork profile can prove you deliver for real clients, but build a portfolio first, start with small projects, and don’t trade study time for low-paid work.</li></ol><h3>3. Your First 90 Days</h3><ul><li><strong>Week 1:</strong> Start the reflection log. <br><em>Pass mark: three lines written.</em></li><li><strong>Week 2:</strong> Set up your public résumé and add one project README answering the three questions. <br><em>Pass mark: a stranger can run your project.</em></li><li><strong>Week 3:</strong> Choose one channel and one course, and finish the first module. <br><em>Pass mark: notes tagged with “where I can use it.”</em></li><li><strong>Week 4:</strong> Publish one LinkedIn post about what you learned. <br><em>Pass mark: published, not perfect.</em></li><li><strong>Weeks 5–6:</strong> Find a beginner-friendly issue (goodfirstissue.dev, firsttimersonly.com, Hacktoberfest or MLH) and submit a pull request. <em>Pass mark: submitted. Merged is a bonus.</em></li><li><strong>Week 7:</strong> Do one informational interview. <br><em>Pass mark: three things you didn’t know before.</em></li><li><strong>Week 8:</strong> Review your log and choose a certification tied to a project. <br><em>Pass mark: a written reason for the choice.</em></li><li><strong>Weeks 9–10:</strong> Build or extend a project and deploy it. <br><em>Pass mark: a live URL.</em></li><li><strong>Week 11:</strong> Record your three-minute pitch. <br><em>Pass mark: you can deliver it without reading.</em></li><li><strong>Week 12:</strong> Post a “what I learned in 90 days” update. <br><em>Pass mark: published.</em></li></ul><blockquote><em>After 90 days, repeat the cycle. Every six months, take on one more level of responsibility.</em></blockquote><h3>4. Five Common Mistakes (and Quick Fixes)</h3><p><strong>1. Collecting certificates without building anything.</strong><br>A certificate on its own is a claim. A project is evidence. <br><em>Fix:</em> start with the project, then certify to fill a gap you actually hit.</p><p><strong>2. Starting six habits at once.</strong><br>By week three, all six have collapsed and you feel like you’ve failed. <br><em>Fix:</em> one habit, two weeks, then add the next.</p><p><strong>3. Consuming content instead of producing it.</strong><br>Ten tutorials watched feels like progress, but employers can’t see it. <br><em>Fix:</em> for every course, ship one small project and one post.</p><p><strong>4. Posting without feedback.</strong><br>A post nobody critiques, or a project nobody runs, is decoration, not evidence. <br><em>Fix:</em> send each piece to one person more experienced and ask, “What’s the biggest flaw here?”</p><p><strong>5. Waiting until you feel ready.</strong><br>Readiness usually comes after the first awkward attempt, not before it. <br><em>Fix:</em> publish the imperfect version and revise it.</p><h3>5. How to Know It’s Working</h3><p>Measure monthly, not daily. Track leading indicators (posts, pull requests, conversations) and lagging ones (interviews, offers). What gets reviewed gets improved.</p><h3>6. Final Word</h3><p>This isn’t a checklist to finish. It’s a system to run. Nobody does all six habits well in the same month. What matters is that each month, one thing you did is <strong>visible, verifiable, and slightly outside your comfort zone.</strong></p><p>Start with the reflection log this Sunday. Everything else follows from there.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=72d716e05b12" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>