<?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 Pranav Pathak on Medium]]></title>
        <description><![CDATA[Stories by Pranav Pathak on Medium]]></description>
        <link>https://medium.com/@pranavVersed?source=rss-c3ab16426591------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*B5v08aBLe6I2VZJ9U288bA.jpeg</url>
            <title>Stories by Pranav Pathak on Medium</title>
            <link>https://medium.com/@pranavVersed?source=rss-c3ab16426591------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 14:43:48 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@pranavVersed/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[Metrics and the inefficiencies of language]]></title>
            <link>https://medium.productcoalition.com/metrics-and-the-inefficiencies-of-language-fb90206750ad?source=rss-c3ab16426591------2</link>
            <guid isPermaLink="false">https://medium.com/p/fb90206750ad</guid>
            <category><![CDATA[kpi]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[metrics]]></category>
            <dc:creator><![CDATA[Pranav Pathak]]></dc:creator>
            <pubDate>Tue, 22 May 2018 22:48:04 GMT</pubDate>
            <atom:updated>2018-05-30T12:41:18.521Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/375/1*9Mf-LFR_CW_3eL1XGAAbFg.png" /></figure><p>Being a Product Manager makes me question some fundamental constructs of language that we take for granted. A few cases in point:</p><ol><li>Guests should feel safe in a hotel — Define ‘safe’</li><li>Can we trust this customer? — Define ‘trust’</li><li>Is the website easy to use? — Define ‘easy’</li><li>Is this good? — Define ‘good’</li></ol><h4>The goopy, wobbly, abstract mass of ‘language’</h4><p>Languages often fail to communicate quantity. We rely on mental maps and past experiences to draw quantities from abstraction. How does ‘safe’ make you feel? Try to quantify that feeling. Think ‘happy’ — explain how happy you are. We rarely ever use numbers to quantify these concepts. Languages by themselves are aware of this abstraction, giving us a superlative to define these feelings relatively — you can be happy, very happy, happiest. Good, Great, Awesome!</p><p>Metrics cannot measure these abstract concepts and their relative degrees, and that is a severe constraint for Product Development. Success cannot be measured with abstract goals. Have we really built a better experience than we had? Will users be more likely to use our product now? Is the product intuitive? We struggle to quantify product success because of these abstractions.</p><h4>We build metrics to overcome the intrinsic inefficiencies of abstraction</h4><p>Placeholder indicators and metrics such as conversion, customer lifetime value, repeat purchase rate, frequency of interaction, and dozens of other metrics try to force fit numbers into meaning. Metrics can work great, as long as the full context surrounding them is well understood.</p><p>But torture numbers, and they confess to anything — anything being the keyword. Our limited understanding of the context that comes with metrics, coupled with our cognitive biases, has led us to dark UX patterns. No one ever said “let’s trick people into using our products”. Regardless, happenings like this are fairly common:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*PPPV_o4KceMLlt3nzdrcpA.jpeg" /></figure><p>Apparently, you must install an app to read the news. The goal was to increase app downloads, and this strategy worked. The fact that this may not be a good customer experience stays in the shadows because ‘good’ is not a measurable quantity. There are workarounds of course. For example, monitoring retention rates, uninstall rates, time spent on app, the whole shebang, but these metrics don’t come without their own pitfalls.</p><p>Besides, pre-defined success metrics enhance our biases, and can lead to decisions that may be great for short term success, but detrimental to the product in the long run. Working with the goal of increasing app installs, one would expect certain other metrics to be negatively impacted. Our inherent biases would make us ignore this impact because we are building in pursuit of a single north star metric.</p><h4>Let’s build metrics to solve problems, and build features to optimize for metrics.</h4><p>While it is imperative to define success metrics, the optimization of metrics should not be the goal of product design. Coming up with product ideas to manipulate metrics is fairly simple. Just about anyone can ideate to optimize metrics, no product skill needed. Rules can be defined to ideate on and run UI optimizations. We are very close to having ML models churn out infinite hypothesis and validate them based on metrics. The larger your user base, the easier it is to manipulate metrics.</p><p>Product Managers do not add value by making hypothesis to manipulate metrics. Rather, they bring to the table an understanding of the abstract user problem, and the consequences to a proposed solution, both tangible and intangible. With this in mind, it is a bad idea to make product decisions based on the metrics they impact. Rather, metrics should be used as an indicator of success for a product built to solve for a user need.</p><p>A good hypothesis targets solving for a user problem — ease of transaction, reduced friction, superior experience, safety, loyalty. The aim is a concrete definition of a user need, followed by brainstorming for potential solutions. Once a few solutions are on the table, a concrete hypothesis needs to be formed. Only then should the metrics to target be discussed.</p><p>For example, the problem could be to reduce friction and simplify the buying experience for an e-commerce platform. For this, we want to enable the user to filter by their favorite brands. The hypothesis is that by enabling the user to filter products by their favorite brand, we enable them to find a product they like. The metrics which should be used to measure this experience, then, flow naturally. For the cohort of users who are exposed to the new filter, we can monitor feature usage (users setting a brand filter), time spent searching for the product (indicator of finding the product faster because of the filter), conversion rate (indicator of finding the product that the user wants), return rates (indicator of satisfaction with the final product). Of course, this is an oversimplification. The nuances to metrics for e-commerce are far more complex, but the approach illustrates the point.</p><p><em>Let’s build products to solve customer pain points, no matter how abstract they seem. Let’s be proactive in trying to quantify these abstractions, and the metrics we need to influence will follow naturally.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fb90206750ad" width="1" height="1" alt=""><hr><p><a href="https://medium.productcoalition.com/metrics-and-the-inefficiencies-of-language-fb90206750ad">Metrics and the inefficiencies of language</a> was originally published in <a href="https://medium.productcoalition.com/">Product Coalition</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A Product Management Framework for Machine Learning — Part 3 of 3]]></title>
            <link>https://medium.productcoalition.com/a-product-management-framework-for-machine-learning-part-3-of-3-2622a129595c?source=rss-c3ab16426591------2</link>
            <guid isPermaLink="false">https://medium.com/p/2622a129595c</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Pranav Pathak]]></dc:creator>
            <pubDate>Tue, 17 Apr 2018 02:15:38 GMT</pubDate>
            <atom:updated>2018-05-30T14:23:33.862Z</atom:updated>
            <content:encoded><![CDATA[<h3>A Product Management Framework for Machine Learning — Part 3 of 3</h3><p>This is part three of a series of 3 posts — if you haven’t yet, here’s the link to read <a href="https://productcoalition.com/my-framework-for-machine-learning-products-part-1-of-3-2354a0da63d2">part 1</a>, and here for <a href="https://productcoalition.com/a-product-management-framework-for-machine-learning-part-2-of-3-399e82609457">part 2</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/574/1*iMTU6eb17zlecQZ60FqQvQ.png" /></figure><p>For the final installment of this series, we discuss monitoring, and how Product Managers can add value to Machine Learning projects.</p><h3>Monitor for Errors, Technical Glitches, and Model Accuracy</h3><p>If you’ve ever built a successful and sustainable product, you’ve taken monitoring seriously. A quick run through why monitoring is important, especially in the context of ML systems:</p><h4>Why do you need to monitor?</h4><ol><li><strong>You’ve built a complex system with multiple moving parts</strong><br>Machine Learning products are complex and evolving. They rely on UX and user journey maps. One Product Manager does not necessarily control all parts of this experience. Imagine for example another team removing your ‘thumbs up’ feedback loop to optimize for conversion. This can quickly snowball into bad retraining data for your model, and in some cases, render the model useless.</li><li><strong>Radical changes to customer behavior need to be captured </strong><br>Customer preferences, wants, and tastes change. Business cycles change. This means that the model is under continuous stress to learn and adapt. New features may need to be added, old features may need to be removed, and changes to customer behavior need to be captured — all to ensure that model performance does not degrade over time.</li><li><strong>Models need to be retrained</strong><br>Even without any external influence, models lose their calibration over time and the performance can continuously degrade. Having good monitoring ensures retraining with new constraints well within time.</li></ol><p>The metrics for monitoring performance are usually radically different from business metrics. Monitoring for something link conversion, and setting an alert every time conversion tanks simply does not cut it. Here are the reasons why:</p><ol><li><strong>Business metrics lag</strong><br>When monitoring from business metrics, what you see is an effect and not a cause — conversion monitoring, for example, indicates only that customers are not converting. This could be due to multiple factors within the user journey, seasonality, business cycles, new launches, just about anything. Business metrics are always a measure of reaction from customers. They say nothing about system health, and usually need time to mature. A glitch in the e-commerce delivery system, for example, will only reflect in the product closer to the customer’s delivery time. This could mean days of lag before you even discover that there is a glitch, all the while losing precious customers. You want a monitor that can detect potential system issues in time and mitigate them before problems creep up.</li><li><strong>You (almost) never have a real time view of business metrics</strong><br>Even if you choose to monitor a specific business goal — clicks on the ‘buy’ button for example — business metrics will still lag. This is mainly because of calculation overheads. Data processors responsible for business KPIs perform extensive calculations and process large amounts of data. The consequence is that often, business dashboards lag by a few hours if not more (If you can see your business metrics realtime, I am deeply and frighteningly jealous).</li><li><strong>It’s a nightmare to debug business metrics</strong><br>So your conversion tanked. What now? Let’s go thought 6,000 lines of code, and look at all the wrong predictions that your model spilled out — not very scalable. System and model specific monitors are great at pointing to the exact location of the bug that can lead to potential catastrophes. Debugging is easier with good alerting and affords precious time with rollbacks.</li></ol><h4>What should you monitor?</h4><p>For the purpose of this discussion, I will not discuss monitoring for the non-ML bits. However, I am taking a giant leap of faith and assuming that bugs in code are already being monitored — never, ever rule out that part. The Machine Learning model itself should be monitored for precision and recall. Set up an acceptable threshold, and set an alert every time the model’s performance breaches that threshold.</p><p>I usually set aside x% of my traffic for perpetual monitoring. For example, if you’re building a recommender system, it’s always a good idea to have a sample set that never interacts with the recommender system and monitor their KPIs. If you’re automating human processes, always send a sample of the model’s predictions to a human agent to validate. A good example is Spam detection — if 100% of your traffic goes through an ML spam detector, you never know if the model is right — you work under the presumption that everything marked by the model is legitimately spam. Having a human agent validating this assumption and monitoring the precision gives an added confidence that all is good with the ML system.</p><p>All Machine Learning models are in essence precise probability calculators. Monitoring ensures that the system is healthy and reliable at all times.</p><h3>Get Creative!</h3><blockquote>Or why Product Managers are essential to ML success</blockquote><p>As can be seen throughout the framework, Product Managers play a role in the entire ML System journey right from conceptualization to production and beyond. PMs are deeply aware of the customer journey, the underlying business logic, the happy paths and the disasters. As someone who sees the full picture, PMs can not only influence the ML Product, but also become the difference between success and failure.</p><p>For some final words, here is a quick summary of the framework, and how you as a PM can add value.</p><p><strong>1 . Identify the problem<br></strong>Product Managers are a strong voice of the user. Identifying the right problem, with great business and customer impact, is key to a delightful ML Product.</p><p><strong>2 . Get the right data set <br></strong>Machine learning needs data. Think about creative ways to generate reliable data sets — ask your users, interpret their preferences from their actions, ask humans to generate the data set, or use a combination of all three! Facebook has at least one Product guy who is super happy about his data collection approach:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qDX6c32JHya6t8hJR7aZkA.gif" /></figure><p>Facebook reactions are awesome — they leverage peer to peer engagement to determine the popularity and sentiment of a post, and can be used to curate the user’s feed.</p><p><strong>3 . Fake it first<br></strong>This one simply cannot work without a little Product magic. Think out of the box to conjure an experience that mimics the one you want to build. Here’s an extreme example:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*JlzLjfAnlf7F8qZUlmioew.gif" /></figure><p>Ford dressed up one of their employees as a car seat to simulate self driving cars and study the reactions of everyone on the road. Here’s the full <a href="https://www.theguardian.com/technology/2017/sep/15/self-driving-car-fake-ford-virginia-tech-man-in-seat">article</a>.</p><p><strong>4 . Weigh the cost of getting it wrong<br></strong>It’s your job to protect your users and the business. Ensure awareness of all the pitfalls of Machine Learning. Remember, you are what stands between your users and this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*xPA_6XKn5xp-up7Cof5eSw.jpeg" /></figure><p><strong>5 . Build a safety net</strong><br>You know the disaster scenarios, and you know how to get your users out of them. Ensure that users are not stuck in loops or spiraling down disasters waiting to happen. Never, ever do what this guy did:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/777/1*AfJ7QMtyVcKMEJjjIQMaMw.png" /></figure><p><strong>6 . Build a feedback loop — </strong>This helps to gauge customer satisfaction and generates data for improving your machine learning model. We’ve covered this extensively in <a href="https://productcoalition.com/a-product-management-framework-for-machine-learning-part-2-of-3-399e82609457">Part 2</a>. Here’s something funny instead:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/400/1*50UG1NURnJ7mpRch5MqkUA.jpeg" /></figure><p><strong>7 . Monitor — </strong>For errors, technical glitches and prediction accuracy.</p><p>And for some parting words — Data Scientists can build amazing models, but Product Managers translate them into delightful and usable products. Don’t let the jargon intimidate you — ML models are, simply put, precise probability calculators.</p><p>That concludes the series. Let me know (here or in a message) if this helped, or if you have any questions. Let me also know what I could write next!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2622a129595c" width="1" height="1" alt=""><hr><p><a href="https://medium.productcoalition.com/a-product-management-framework-for-machine-learning-part-3-of-3-2622a129595c">A Product Management Framework for Machine Learning — Part 3 of 3</a> was originally published in <a href="https://medium.productcoalition.com/">Product Coalition</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A Product Management Framework for Machine Learning — Part 2 of 3]]></title>
            <link>https://medium.productcoalition.com/a-product-management-framework-for-machine-learning-part-2-of-3-399e82609457?source=rss-c3ab16426591------2</link>
            <guid isPermaLink="false">https://medium.com/p/399e82609457</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Pranav Pathak]]></dc:creator>
            <pubDate>Tue, 27 Mar 2018 21:30:29 GMT</pubDate>
            <atom:updated>2018-03-27T21:30:29.490Z</atom:updated>
            <content:encoded><![CDATA[<h3>A Product Management Framework for Machine Learning — Part 2 of 3</h3><p>This is part two of a series of 3 posts — if you haven’t yet, here’s the<a href="https://productcoalition.com/my-framework-for-machine-learning-products-part-1-of-3-2354a0da63d2"> link to read part 1</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*x_oHroISu3kM_s6mkkVpKQ.png" /></figure><p>A quick look-back at the 8 steps to building an AI Product:</p><ol><li><strong>Identify the problem</strong><br>There are no alternatives to good old fashioned user research</li><li><strong>Get the right data set</strong><br>Machine learning needs data — lots of it!</li><li><strong>Fake it first</strong><br>Building a Machine Learning model is expensive. Try to get success signals early on.</li><li><strong>Weigh the cost of getting it wrong</strong><br>A wrong prediction can have consequences ranging from mild annoyance to the user to losing a customer forever.</li><li><strong>Build a safety net</strong><br>Think about mitigating catastrophes — what happens if Machine Learning gives a wrong prediction?</li><li><strong>Build a feedback loop</strong><br>This helps to gauge customer satisfaction and generates data for improving your machine learning model</li><li><strong>Monitor</strong><br>For errors, technical glitches and prediction accuracy</li><li><strong>Get creative!</strong><br>ML is a creative process, and Product Managers can add value</li></ol><p>This article covers steps 4 through 6. Let’s get right to it!</p><h3>4. Weigh the cost of getting it wrong</h3><p>Machine Learning is not error free, and developing without guardrails can have serious consequences. Consider for example <a href="https://www.theverge.com/2016/3/24/11297050/tay-microsoft-chatbot-racist">this</a> malicious twitter bot:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1008/1*b8X8kPkC0KCYJgegNLS6ZQ.png" /></figure><p>Or <a href="https://www.wired.com/story/when-it-comes-to-gorillas-google-photos-remains-blind/">this</a> Google Photos AI gone wrong.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/580/1*DEoCT6MsHrnshG2_PCwoeQ.png" /></figure><p>Of course, we cannot conclude that all AI will go insane. At the same time, we need to acknowledge that there is a cost to getting it wrong with AI. Think for example an order management service that automatically detects if a customer wants to cancel their order. The cost here is that the model might wrongly interpret the user’s intent as a cancellation, thereby having financial consequences for the user and the company.</p><p>Machine learning relies to a large extent on probabilities. This means that there is a chance that the model gives the wrong output. Product Managers are responsible for anticipating and recognizing the consequences of a wrong prediction. One great way to anticipate consequences like the ones above is to test, test, and test some more. Try to cover all scenarios that the AI might have to encounter. Understand what makes up the probabilities computed by the model. Think about the desired precision of the model (keep in mind, the more precise your model is, the lesser the cases it can cover — more on this in part 3). Talk to your Data Scientists about what can potentially go wrong. Ask tough questions — it is the Data Scientist’s job to build a fairly accurate model, and it is the Product Manager’s job to understand the consequences of such a model.</p><p>In general, the cost of getting it wrong varies by use case. If you are building a recommender system to suggest similar products, where you previously had nothing, the worst outcome might be a low conversion on the recommendations you offer. One must then think about how to improve the model, but the consequences of getting it wrong are not catastrophic.</p><p>If AI is automating manual flows, a good estimate for cost is how many times an AI gets it wrong in comparison to humans. If for example, a human is right about an email being Spam 99% of the time, an AI Spam detector is worth the investment at a 99.1% precision.</p><p>It is important to understand these consequences in order to mitigate them and to train the model better for future scenarios.</p><h3>5. Build safety nets</h3><p>Once all of the consequences of getting a prediction wrong are identified, relevant safety nets must be built to mitigate them. Safety nets can be intrinsic and extrinsic.</p><p><strong>Intrinsic safety nets</strong></p><p>These recognize the impossibilities that are fundamental to the nature of the product. For example, a user cannot cancel an order if there is no order made in the first place. So that email you got is definitely not talking about cancelling an order, and the model has misclassified. It’s a good idea to have a human agent look into this case. A useful activity is to map out the user journey for your product, and identify the states that the user can go through. This helps weed out impossible predictions. Intrinsic safety nets are invisible to the user.</p><p><strong>Extrinsic safety nets</strong></p><p>Extrinsic safety nets are visible to the user. They can take the form of confirming user intent or double checking the potential outcome. LinkedIn for example has a model to detect the intent of a message and suggest replies to its users. It does not, however, assume a reply and automatically send it. Instead, it asks the user to pick form a list of potential replies.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ESSU43osFE5JF6R3gpD2jA.png" /></figure><p>Extrinsic safety nets for users are not a new concept. Think about every time your Windows 95 popped up this dialog box:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/576/1*4fnd_utiywNUVfxAPEgbfg.jpeg" /></figure><p>This system does not use AI, but does take into account that erroneous actions can have consequences. Safety nets are ingrained in all products, and AI is not an exception.</p><h3><strong>6. Build a feedback loop</strong></h3><p>Setting up safety nets also helps gather the much needed feedback for the model.</p><p>Feedback loops help measure the impact of a model and can add to the general understanding of usability. In the context of an AI system, feedbacks are also important for a model to learn and become better. Feedbacks are an important data collection mechanism — they yield labelled datasets that can be directly plugged into the learning mechanism.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6Whc4u6f1NfGbICozinBLQ.png" /></figure><p>For Amazon’s recommendation module, the feedback loop is quite simple — does the guest click on the recommendations, and is this recommendation bought?</p><p>AirBnB uses a more direct approach for feedback collection.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DVUBeaMCaw-e541eWHsV5Q.png" /></figure><p>Netflix uses a hybrid. It can understand based on how many of the recommendations you click and view, and also uses a thumbs up mechanism for explicitly logging preferences.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1_XJEbBynDWbe5KHkSkdow.png" /></figure><p>It must be noted that safety nets can often double down as feedback channels for refining the model. Safety nets by their nature are outside the ambit of the model’s prediction. They should be used, whenever possible, to label data and generate a stronger learning dataset.</p><p>I make AI sound scary, but responsible development while understanding consequences is essential to any product regardless of whether or not AI is involved.</p><p>Just to reiterate, <a href="https://productcoalition.com/my-framework-for-machine-learning-products-part-1-of-3-2354a0da63d2">here is the link to the first post in this series</a>. Make sure to let me know if this was helpful!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=399e82609457" width="1" height="1" alt=""><hr><p><a href="https://medium.productcoalition.com/a-product-management-framework-for-machine-learning-part-2-of-3-399e82609457">A Product Management Framework for Machine Learning — Part 2 of 3</a> was originally published in <a href="https://medium.productcoalition.com/">Product Coalition</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A framework for Machine Learning Products — Part 1 of 3]]></title>
            <link>https://medium.productcoalition.com/my-framework-for-machine-learning-products-part-1-of-3-2354a0da63d2?source=rss-c3ab16426591------2</link>
            <guid isPermaLink="false">https://medium.com/p/2354a0da63d2</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[product-development]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Pranav Pathak]]></dc:creator>
            <pubDate>Mon, 05 Mar 2018 22:30:45 GMT</pubDate>
            <atom:updated>2018-03-29T15:44:25.861Z</atom:updated>
            <content:encoded><![CDATA[<h3>Product Thinking for Machine Learning — Part 1 of 3</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*6_MfuVNhnecVLT1wqyq8hA.png" /></figure><p>Building your first Machine Learning product can be overwhelming — the sheer amount of moving parts, and the number of things that need to work together can be challenging. I’ve often seen great Machine Learning models fail to become great Products, not because of the ML itself, but because of the supporting product environment. UX, Processes, and Data, all contribute to the success of a Machine Learning model.</p><p>Over the years of building Machine Learning products, I’ve come up with a framework that usually works for me. I break down a Machine Learning product into eight steps. Here they are at a glance:</p><ol><li><strong>Identify the problem</strong><br>There are no alternatives to good old fashioned user research</li><li><strong>Get the right data set</strong><br>Machine learning needs data — lots of it!</li><li><strong>Fake it first</strong><br>Building a Machine Learning model is expensive. Try to get success signals early on.</li><li><strong>Weigh the cost of getting it wrong</strong><br>A wrong prediction can have consequences ranging from mild annoyance to the user to losing a customer forever.</li><li><strong>Build a safety net</strong><br>Think about mitigating catastrophes — what happens if Machine Learning gives a wrong prediction?</li><li><strong>Build a feedback loop</strong><br>This helps to gauge customer satisfaction and generates data for improving your machine learning model</li><li><strong>Monitor</strong><br>For errors, technical glitches and prediction accuracy</li><li><strong>Get creative!</strong><br>ML is a creative process, and Product Managers can add value</li></ol><p>For the sake of my sanity and for better readability, I’ll split this post into three parts. This part covers steps 1 through 3. Here goes step 1!</p><h3>1. Identify the Problem</h3><p>Machine Learning is a powerful tool for solving customer problems, but it does not tell you what problems to solve. Before even beginning to decide on whether Machine Learning is the right approach, it is important to define the problem.</p><p><strong>Invest in user research</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*knM96vJfvzGiGgHpobA8Ng.png" /></figure><p>There are no shortcuts to good old fashioned user research. Every successful product identifies a specific user need and solves for it. Conduct a thorough user research in order to identify the pain points of the user and prioritize them according to user needs. This helps to build a user journey map, identifying critical flows and potential roadblocks. Further, the roadmap is super useful for defining processes and flows that need to be modified for your ML solution to work in the first place. <a href="https://medium.com/design-ibm/the-vital-role-of-user-research-8c2f51a9cead">Here</a> is a summary of what good user research should accomplish.</p><p>Once the problem is identified, we need to explore if machine learning offers the best solution.</p><p><strong>Look for one of these characteristics in your problem:</strong></p><ul><li><em>Customization or personalization problem — one size does not fit all</em><br>Problems where users with unique characteristics need to be identified are usually great candidates for machine learning.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7hLGdO-xqBiHyM5gBKAw2w.png" /><figcaption>Amazon’s recommender system identifies unique users who would be interested related products</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/618/1*i_BK42VyHgUws6DPvE7nZA.png" /><figcaption>Netflix uses a ‘match percentage’ to identify shows that most closely match your preferences</figcaption></figure><p>Personalization or customization problems look to identify specific users and cohorts that may be interested in specific content. They are usually driven by past user actions, user demographics, etc.</p><ul><li><em>Repeatable sequence of steps</em><br>Processes which require a repetition of the same sequence of steps once the problem is identified are usually great candidates for automation. For example, the entire ‘cancel your order’ flow in e-commerce can be automated if the intent of automation is identified. Further, recommendations for next steps can also be made based on a user’s last action.</li><li><em>Recognizes or matches pattern</em><br>Look for repeated patterns that you can learn from. Spam engines, for example, index characteristics of spam messages (including text, subject line, sender information, etc.) and look at how many users marked a similar message as spam in order to to identify a potentially spam email.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UDaKf7byMc41ywQQvfCjaw.gif" /><figcaption>Google’s Spam engine marks messages like these based on previous data and message characteristics</figcaption></figure><p><a href="https://eng.uber.com/cota/">Uber’s approach to customer service</a> relies on identifying the best solution for a specific problem based on previous actions of customer service agents.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2E3ozoAyZnT4CM7Ly2QGVg.png" /><figcaption>Learning from humans for a better experience.</figcaption></figure><h3>2. Get the right data set</h3><p>The success or failure of machine learning relies on the coverage and quality of your data set. A good data set has two characteristics — a comprehensive feature list, and an accurate label.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kq6Hshgip5-HKoYWZ9Cgbw.png" /><figcaption>Google Draw’s open datasets are awesome!</figcaption></figure><p><strong>Features</strong> are the input variables to your model. For example, while building a recommendation system, you may want to look at a user’s purchase history, the closest matches for the products they bought, their buying frequency, and so on. Having a comprehensive feature list ensures that the ML model understands enough about the user in order to make a decision.</p><p><strong>Labels</strong> tell the model the right from the wrong. A machine learning model is trained iteratively on a data set. In each iteration, the model makes a prediction, checks if the prediction is right, and calibrates itself for wrong predictions. It is thus important for the model to know whether or not a prediction is right. Labels convey this information to the model. For a recommender system, this label will be whether the recommended item was indeed purchased. Having the right set of labels influences the model’s performance.</p><p><strong>Bonus</strong>: Some awesome resources to get you started into what labelled data looks like:<br>* <a href="https://www.kaggle.com/datasets">Kaggle datasets</a> : Huge collection of labelled data on just about anything!<br>* <a href="https://quickdraw.withgoogle.com/data/">Quickdraw</a>: Crowdsourced data on line images and strokes</p><h3>3. Fake it First</h3><p>Investing in machine learning models is expensive. Building a dataset with the right features and labels, training the model and putting it in production can range from a few weeks to a few months. It is important to get a signal early on and validate if the model will work. A good idea is to fake the interaction first, for a small set of users.</p><p>For personalization, have a list of items ready for a user to select, based on what they selected last. Simple rule-based engines are often the first steps to evolution into a more complex machine learning model. Some examples of rule-based recommenders:</p><ul><li>If the user bought pasta, they probably need cheese too</li><li>A user buys toothpaste once a month, so we recommend it to them a month after their last purchase</li><li>A user always pays by credit card, so we surface that as the default payment option next time</li></ul><p>It’s obvious that we cannot write simple rules to cover every case — and that is when the power of machine learning can be used best — but a few simple rule based proxies go a long way into validating the outcome of the machine learning approach.</p><p>For automating customer flows, it may be a good idea to ask a human on the other side to do exactly what an automation would do — and nothing else. Companies often test if they should build chatbot through a controlled release with a human answering questions on the other side.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5bxYB2E97da3GO5Zs_A9nw.png" /></figure><p>The idea is to test if users respond positively to machine learning. While these techniques might not give the best results, they are important for getting a signal. Getting a signal early on can save time and effort, and help correct the vision and direction of the product. This is your best shot at guaranteeing returns on the investment put into building a Machine Learning system.</p><p><a href="https://productcoalition.com/a-product-management-framework-for-machine-learning-part-2-of-3-399e82609457">Here is the link to part 2.</a></p><p>Phew! Writing that was more work than I anticipated. Let me know if this was useful so far, and what more you would like to hear. I’ll keep writing more about the approach, and your feedback will make future posts better.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2354a0da63d2" width="1" height="1" alt=""><hr><p><a href="https://medium.productcoalition.com/my-framework-for-machine-learning-products-part-1-of-3-2354a0da63d2">A framework for Machine Learning Products — Part 1 of 3</a> was originally published in <a href="https://medium.productcoalition.com/">Product Coalition</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What I learnt about Product Management from a 1000 year old city, a man in chains, and Customer…]]></title>
            <link>https://medium.productcoalition.com/what-i-learnt-about-product-management-from-a-1000-year-old-city-a-man-in-chains-and-customer-83c6b4751619?source=rss-c3ab16426591------2</link>
            <guid isPermaLink="false">https://medium.com/p/83c6b4751619</guid>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[customer-experience]]></category>
            <category><![CDATA[amsterdam]]></category>
            <category><![CDATA[customer-service]]></category>
            <dc:creator><![CDATA[Pranav Pathak]]></dc:creator>
            <pubDate>Wed, 27 Dec 2017 13:09:13 GMT</pubDate>
            <atom:updated>2017-12-27T13:29:49.160Z</atom:updated>
            <content:encoded><![CDATA[<h3>What I learnt about Product Management from a 1000 year old city, a man in chains, and Customer Service Agents</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/768/1*l3n3_ytyySEDFUbkY3I0jw.jpeg" /></figure><p>I moved to Amsterdam 7 months ago. I wanted to get out of my comfort zone. I had built for India, and now I wanted to build for a different audience.</p><p>I moved to a new country in the same year that I crossed Indian shores for the first time in my life. I started building for ten times as many people as I had ever built for, scattered across continents, speaking different languages. I expected radical changes in my life and my perspectives.</p><p>I expected a culture shock. I didn’t get one. Between sips of chai with a Brazilian colleague, discovering the striking parallels between Indian and Kazakh culture over a late night coffee, and a Drag Queen Bingo Night, I was easily assimilated into the cosmopolitan that Amsterdam is. Making friends is the same everywhere.</p><p>For work, I listened to customer calls for a large part of a few days. Customer Service calls are a powerful window into your audience. You would only call customer service at the height of emotion. When your’e overjoyed or thankful, or for most part, when you’re frustrated, is when you would pick up the phone. I listen to calls from all over the world. And each of these calls share the same messages. The same concerns, echoed day and night, in 48 languages, in 24 countries. Our concerns are the same everywhere.</p><p>On a quaint evening in Dam Square, a group had formed around the center, and suspecting a street performance, I huddled in. What I saw was a man attempting Houdini’s Great Escape. I seemed to have reached at the right moment. He had just been bound in chains, and standing on his makeshift stage, he spoke</p><blockquote><em>Before I attempt to escape these chains and disappear, look around you. You will see every colour that we’re born with. There is someone around you from every continent. All of you are standing here, watching me, getting amused and entertained, with smiles on your faces. That is the reason I perform — street art has the power to unify, and to bring joy and peace.</em></blockquote><p>Delight is the same everywhere…</p><p>I was out to understand different users. I had built for India, I wanted to build for a different audience. Amsterdam taught me that my audience is the same everywhere.</p><p>If you enjoyed reading this story, clap, comment or share! It encourages me to keep writing.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=83c6b4751619" width="1" height="1" alt=""><hr><p><a href="https://medium.productcoalition.com/what-i-learnt-about-product-management-from-a-1000-year-old-city-a-man-in-chains-and-customer-83c6b4751619">What I learnt about Product Management from a 1000 year old city, a man in chains, and Customer…</a> was originally published in <a href="https://medium.productcoalition.com/">Product Coalition</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[My last memory of my Grandfather]]></title>
            <link>https://medium.com/@pranavVersed/my-last-memory-of-my-grandfather-107cd0bfb017?source=rss-c3ab16426591------2</link>
            <guid isPermaLink="false">https://medium.com/p/107cd0bfb017</guid>
            <category><![CDATA[memories]]></category>
            <category><![CDATA[short-story]]></category>
            <category><![CDATA[college]]></category>
            <dc:creator><![CDATA[Pranav Pathak]]></dc:creator>
            <pubDate>Wed, 27 Dec 2017 12:44:18 GMT</pubDate>
            <atom:updated>2017-12-27T13:07:41.156Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/382/1*5igAVmlzU37GzH-Ki_K-Rg.png" /></figure><p>Rummaging through some preserved photos on an old phone, I came across this — my grandfather in his thirties, with a head full of hair and a thick black moustache. I’ve never seen him so young in my life. I was told he was dean at the local college, and I imagine him shaping the lives of his students in a stern, authoritative and distant fashion. But that is not how I remember him. In fact, my most vivid memory of him is quite the opposite.</p><p>I was back home from college in the summers. My grandfather was home too. Lazy summer afternoon, I was watching something on the telly. He called me to his room. Reluctance — it took me all of my willpower to get out of the couch and onto the next room, ten steps away. Why wasn’t he asleep? Why does he need to disturb my peace?</p><p>The curtains were pulled close. The room was warm, the ceiling fan spun full speed. The humidity and the darkness did nothing to absolve me of eternal inertia. He was sitting on a corner of the bed, his trusty cane beside him. I entered and he called me closer. What do you need, nanu? Water? Tea? — He was a big fan of chai.</p><p>But this wasn’t about chai. He asked me to open my palm. I did. He placed his fist over my palm. Out poured dozens of salted peanuts! He smiled a toothless smile. I laughed. Sat on the bed beside him. An afternoon of conversation. My best ever memory of him.</p><p>If you enjoyed the story, clap, share or comment. It encourages me to keep writing!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=107cd0bfb017" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A chat with Jay from the Product Coalition]]></title>
            <link>https://medium.com/@pranavVersed/a-chat-with-jay-from-the-product-coalition-2b206c48a7e3?source=rss-c3ab16426591------2</link>
            <guid isPermaLink="false">https://medium.com/p/2b206c48a7e3</guid>
            <category><![CDATA[product]]></category>
            <category><![CDATA[india]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Pranav Pathak]]></dc:creator>
            <pubDate>Sat, 10 Dec 2016 05:07:32 GMT</pubDate>
            <atom:updated>2016-12-10T05:07:32.888Z</atom:updated>
            <content:encoded><![CDATA[<p>Jay from the Product Coalition and I discussed building for emerging markets, the challenges and the triumphs. Watch the video here:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FqzVHxxQ6zBw%3Fstart%3D645%26feature%3Doembed%26start%3D645&amp;url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DqzVHxxQ6zBw&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FqzVHxxQ6zBw%2Fhqdefault.jpg&amp;key=d04bfffea46d4aeda930ec88cc64b87c&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/41355bdc595d4957968b7b7408b9fdf6/href">https://medium.com/media/41355bdc595d4957968b7b7408b9fdf6/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2b206c48a7e3" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[15 minute product design challenge #1: Rating & Review System for Ecommerce]]></title>
            <link>https://medium.com/@pranavVersed/15-minute-product-design-challenge-1-rating-review-system-for-ecommerce-342293362efe?source=rss-c3ab16426591------2</link>
            <guid isPermaLink="false">https://medium.com/p/342293362efe</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[product-development]]></category>
            <category><![CDATA[product-design]]></category>
            <category><![CDATA[ecommerce]]></category>
            <dc:creator><![CDATA[Pranav Pathak]]></dc:creator>
            <pubDate>Sun, 13 Nov 2016 18:34:54 GMT</pubDate>
            <atom:updated>2018-01-12T15:49:52.727Z</atom:updated>
            <content:encoded><![CDATA[<p>A friend asked me to design an MVP product rating and review system for an ecommerce company — within 15 minutes.</p><p>Start time: 11:31pm. End time: 11:47 pm (1 minute to find a pen and paper).</p><p>Here is the result:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*37NXEZ5KkWiUEPP_bWd2gQ.png" /><figcaption>Review system designed in 15 minutes. Red indicates comments.</figcaption></figure><p>I think this works as an MVP quite well. Certain obvious things are missing — language processor and sentiment analysis are long shots (Phase 2 maybe?) The input/output wireframes are pretty basic but solve the purpose. What else can one do in 15 minutes? The typing and sketching took a decent amount of time.</p><p>Very eager for feedback. Let me know what you think! If you have any suggestions for my next 15 minute product design exercise, hit me up.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=342293362efe" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Product Review #1: Hardbound]]></title>
            <link>https://medium.com/@pranavVersed/product-review-1-hardbound-d2f324ef604e?source=rss-c3ab16426591------2</link>
            <guid isPermaLink="false">https://medium.com/p/d2f324ef604e</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[medium]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[product-design]]></category>
            <dc:creator><![CDATA[Pranav Pathak]]></dc:creator>
            <pubDate>Mon, 19 Sep 2016 20:39:05 GMT</pubDate>
            <atom:updated>2016-09-20T16:43:39.943Z</atom:updated>
            <content:encoded><![CDATA[<p>I’ve finally decided to stop procrastinating and write my first product review ever. Proudly presenting what I thought about <a href="https://itunes.apple.com/us/podcast/hardbound/id1120232683?mt=2">Hardbound</a>!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/400/1*6bdmax6MNrj1czYos7_NXA.jpeg" /><figcaption>Hardbound: Stories for curious minds</figcaption></figure><p>Hardbound is a digital storytelling app. The founder’s inspiration was to create online media that, rather than deriving from print, could use the full power of your device to create quality content. Did they succeed? Definitely! Hardbound serves content in visual, interactive snippets like these:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/584/1*w_nfTsIweVFWbzqUb4hbIA.png" /></figure><p>The story flows as you keep swiping. The content is the focus. Short, crisp and to the point. The animation and the flow of the story make even seemingly boring content engaging. The visual feedback is great at capturing attention. I enjoyed reading everything I touched!</p><p>The content is built in-house, rather than being crowd sourced. This, I presume, is to ensure quality. Besides, creating rich content takes resources, which I believe few users will have. The team promises to keep delivering quality content, to the point of not ‘reducing friction’ for the creators. This reflects on the content churn as well — Hardbound publishes only one story a week. Given the quality of the content though, I don’t mind this speed at all! Besides, the app keeps you wanting for more.</p><p>Overall, a fun app. I think I’ll keep it!</p><p><strong>What I liked</strong></p><ol><li><em>Simple UI<br></em>The menu is just three buttons:</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/622/1*YKbc7DaKQ8t2O0Fyz9e6dA.png" /></figure><p>Not many actions to take, and no learning curve — just tap on a story and read.</p><p>2. <em>No categorization of content<br></em>Everything is shown up-front:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/538/1*HxlkN73DmvgVBn-4eNt0vA.png" /></figure><p>I think this is a positive. It brings focus to the one thing they’ve published over the week. The user does not have to swarm through a sea to find what she likes. There is one story told very well, and the user reads it. No frills. I’d like to see what happens when there is a lot more content though.</p><p>3. <em>‘Share’ is well implemented<br></em>Users can share a story right after they read. The content is shared as a web url, so the reader doesn’t need to have the app. I loved how user centric this is. No pushing for downloads, just delivering great content. The share is non-intrusive, too. It only pops up once you’ve read a story.</p><p>Overall, a lot of thought went into the product. The focus on the UX shows, and so do strategic calls. There’s a story about building habit forming products in Hardbound’s collection, and looks like they know what they’re saying!</p><p><strong>What I didn’t like:</strong></p><ol><li><em>The loading state — this one is confusing.</em> <br>When you tap on a story, it opens an image on the screen, and you can’t interact with it right until most of the story has loaded. Besides, there is no feedback. I tapped on the phone and tried to restart the app a few times before I saw the loader at the bottom of the screen.</li></ol><p><strong>And I’m wondering…</strong></p><ol><li><em>Interesting choice of notifications — text messages! <br></em>The app does not have notification capabilities. I’m wondering if this is by design or is planned as an enhancement. Either way, Hardbound sends you a text notification every time they publish (an option for email is also available). Historically, notifications have better open rates than text messages. Hardbound might be an exception because of the kind of content they provide. I rarely ever give an app permissions to notify me, so it might actually be better to send me a text. I’m wondering how many others out there are like me, though. Really want to know how this works out…</li><li><em>How will their content scale?</em> <br>Will it always be in-house? What happens when there are a few hundred stories? How will they categorise? I’m not a big fan of labels and tags on content. I think the serendipity is lost when I filter by tags. How is the team at hardbound thinking about solving discovery?</li></ol><p><strong>Links</strong></p><p>Edit courtesy <a href="https://medium.com/u/cf93b6576863">Harman Singh</a>: The app is only available on iOS.</p><p>Download Hardbound <a href="https://itunes.apple.com/us/podcast/hardbound/id1120232683?mt=2">here</a><br>Hardbound on <a href="https://twitter.com/hardbound">Twitter</a> and <a href="https://medium.com/@hardbound">Medium</a><br><a href="https://www.producthunt.com/tech/hardbound-3">ProductHunt</a> entry<br><a href="https://medium.com/hardbound-co/introducing-hardbound-10e547b2c67f#.kl50ry20c">Here’s</a> what the founders wrote at launch</p><p><em>PS: This is my first review, so I would really love feedback! Do let me know if this review was helpful, whether you would want to hear more, or any other suggestions you might have.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d2f324ef604e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I’m a crazy capitalist, but…]]></title>
            <link>https://medium.com/@pranavVersed/for-democracy-f146e925ebe2?source=rss-c3ab16426591------2</link>
            <guid isPermaLink="false">https://medium.com/p/f146e925ebe2</guid>
            <category><![CDATA[economics]]></category>
            <category><![CDATA[capitalism]]></category>
            <category><![CDATA[marxism]]></category>
            <dc:creator><![CDATA[Pranav Pathak]]></dc:creator>
            <pubDate>Fri, 28 Aug 2015 10:40:11 GMT</pubDate>
            <atom:updated>2015-08-28T11:02:56.275Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/529/1*mLzX46I7YvVNNVz9HTqZRQ.gif" /></figure><p>I’m accused of being a capitalist — a crazy one at that. Not that there’s anything sane about any of them. And not that I’m a true capitalist either. I have a problem with just one idea — “From each according to his ability, to each according to his need”.</p><p>Don’t get me wrong, I am not challenging Marx’s wisdom. Nor am I claiming he was as shaken in the head as the capitalists. But I do have a bone to pick with those who want to see his idea put to practice today. And here’s why…</p><p>You see, the world as we see it today works on incentives. I’m not saying it’s right. I’m not saying it’s wrong. It just is. And the sooner we accept it, the better. A Professor at a local college conducted a great experiment to prove this. Him and his students were arguing about the same Marxian school of thought I write about. So, in an analogy, he said, let’s average out your grades, and let everyone get this average grade. So everyone got a B. Those who didn’t study were happy. Those who did had to face disappointment. So in the next test, those who didn’t study earlier now studied even lesser — they knew there was someone to back them up. Those who did study earlier, reduced their effort. There simply wasn’t enough incentive. The class performance dropped, and everyone got a C. And it didn’t stop here. At one point, the entire class failed. Sure, there were those who studied for the love of the subject. But they were too few in number to keep the averages up. And this is one flaw would hold in the Marxian argument. There are too few of us who work for the love of the job.</p><p>When Marx made his iconic statement, he was talking about a society where technology would be so advanced that there will minimal need for labor. And labor would be intellectually stimulating, challenging, fulfilling. A eutopia of this sort sounds great. But we’re not quite there yet. We’re not even close.</p><p>So incentives are a necessity in the world today. But that does not justify brutal incentivization either. When Ayn Rand talks about rewarding effort and only effort, she again assumes a very egalitarian society. That’s not true. We all start at very different points in our financial and social arrangement. Someone who is worse off in the beginning would have to work much harder than you and me to accomplish the same goals. And having seen poverty at close quarters, I can assure you this would be a disaster. Take the case of an unskilled laborer. He works hard every day, day and night, and in the end, earns just enough to get by. No significant increase in wages, no promotions, no development. Take your own example for a contrast. Let’s say you work just as hard as the daily wage worker. But you have education by your side. You use your mind along with your physical strength. You grow your skill, and with this growth comes professional, personal and financial development. You have avenues for growth, simply because your parents could afford to send you to school.</p><p>You may argue that this is sheer luck, and that you’re entitled to make the best use of what you have. But this approach won’t work to your benefit in the long run either. What this view would end up doing is, it would widen the gap between the well to do and the poor. And this is bad not just for the poor, but for you as well. Kruger and his colleagues did a detailed analysis here. They found that as the gap begins to increase, the income of the entire economy falls. Growth rates fall, consumptions fall, wealth falls. Even the well to do now see that their standard of living is declining — so much so, that the countries with the largest inequality often have the lowest standard of living.</p><p>Believe it or not, but the poorer sections, or the lower middle class, is the highest consumer. And your consumption does not increase in proportion to your wealth. So when you get richer, a large portion of your money is going to be stored away. This money is not working. It’s not contributing to growth — neither the society’s nor yours. The society is at a disadvantage, and so are you.</p><p>Brutal capitalism is not the solution. We need someone to stabilize, to ensure a framework where the incentivization found abundantly in capitalism would work. We need to ensure that there is enough opportunity for this incentivization, and that everyone gets a chance to this opportunity. We need a stabilizing, allocative and efficient government structure. A structure that, rather than removing incentivization, promotes <em>equal</em> incentivization for all under its purview. Such a structure would allow for mankind’s potential to truly blossom. It would make opportunities available for all, and then reward the one making the best use of the opportunity.</p><p>A democratic government is very well equipped to serve this end. To quote Amartya Sen, ‘Democracy is not a luxury that can wait the arrival of prosperity’. Rather, it is born out of the need and desire for prosperity, the desire for development, the desire to be heard and to be rewarded for effort. Given adequate representation, a democracy can ensure that the needs of most are met with priority. It can ensure that its constituents get a platform on which to build their aspirations. It can ensure collective prosperity through the prosperity of the individual. It has the potential to be truly egalitarian in representation and opportunity. It is, I believe, the perfect form of government. Yet, it needs to be made efficient. The kinks and flaws need to be ironed out. It needs to be molded by the voice of its own people. It needs awareness. It needs education and intellectual participation. It must be shaped through the constant involvement of its own people. Through collective effort, democracy can potentially emerge in its purest form, and bestow upon its subjects individual as well as collective accomplishment, development and prosperity.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f146e925ebe2" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>