<?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 Pei-Ran Yu on Medium]]></title>
        <description><![CDATA[Stories by Pei-Ran Yu on Medium]]></description>
        <link>https://medium.com/@peiran.yuu?source=rss-2e3799f0b15e------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*qAqevFEDGxxNtKMWLBW6Sg.jpeg</url>
            <title>Stories by Pei-Ran Yu on Medium</title>
            <link>https://medium.com/@peiran.yuu?source=rss-2e3799f0b15e------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 00:29:07 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@peiran.yuu/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[Discover strategy options for your influencer marketing in seconds!]]></title>
            <link>https://medium.com/@peiran.yuu/discover-strategy-options-for-your-influencer-marketing-in-seconds-33d303af4bc0?source=rss-2e3799f0b15e------2</link>
            <guid isPermaLink="false">https://medium.com/p/33d303af4bc0</guid>
            <category><![CDATA[monte-carlo]]></category>
            <category><![CDATA[influencer-marketing]]></category>
            <category><![CDATA[growth-analytics]]></category>
            <dc:creator><![CDATA[Pei-Ran Yu]]></dc:creator>
            <pubDate>Wed, 13 Mar 2024 09:33:29 GMT</pubDate>
            <atom:updated>2024-10-11T02:53:11.115Z</atom:updated>
            <content:encoded><![CDATA[<h3>Alex, the Founder of a VR game, and His Marketing Dilemma</h3><p>Meet Alex, a passionate game developer with a dream of creating a mind-blowing VR game that transports players to a hyper-realistic world. This immersive experience features stunning visuals, captivating storytelling, and interactive gameplay, redefining the gaming industry and leaving players in awe of the possibilities of virtual reality. However, effectively marketing this creation in a competitive industry poses a challenge for Alex. Traditional strategies fall short in capturing the excitement and immersion of the VR game.</p><p>But Alex won’t give up easily! He knows that the best way to reach gamers is through the influencers they follow. However, with limited funding and a tight budget, every dollar spent must be carefully planned. With a vast number of gamer influencers out there, Alex feels stuck -</p><p>“Where do I even start??”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*85T7GlorGboqF6mV.png" /></figure><p>There are countless influencers out there, with hundreds of thousands just on YouTube alone. They range from nano-influencers, who have a modest following of less than 10k, to micro-influencers with less than 100k followers. And then there are mega-influencers like PewDiePie and Markiplier, who have millions of followers. Each category comes with its own cost levels as well.</p><p>So — where should Alex start? How does he decide? And how much should he allocate to each person? It is indeed a real struggle to find the right balance.</p><h3>Enter the Influencer ROI calculator, coming to the rescue!</h3><p>The idea is very simple: input numbers, get a chart. The chart shows you how many posts to buy from each influencer.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*mvZpmJTWP2BzTzog.png" /></figure><blockquote><em>👀 Just a heads up that this calculator is specifically useful in certain scenarios. If you want to customize or expand it, make sure to check out the “Routz Notes” section below!</em></blockquote><p>On this chart, each distribution plot represents a range of ROI estimations of a particular strategy. The strategies are defined by the number of posts from different influencer categories. Let’s say Alex has a budget of $10,000 for this influencer marketing test, and he wants to spend at least $8,000 out of this $10,000.</p><p>So, armed with information about the influencers — their total followers, average engagement rates, and whatnot — he simply plugs in the numbers. And ta-da! The calculator reveals the following strategies:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*EL2gNkNoPlJB7g6C.png" /></figure><ul><li>Strategy 1: Buy 1 post from each of the 3 categories, leaving $1,700 from the $10k budget. This strategy yields approximately 11 engagements for each dollar spent.</li><li>Strategy 2: Buy 2 posts from the micro-influencer category and 1 post from a mega influencer, leaving $500 from the $9,500 budget. This strategy results in around 10.5 engagements for each dollar spent.</li><li>Strategy 3: Buy 2 posts from the nano-influencer category and 1 post each from the micro and mega influencer categories. This leaves $900 from the $9,100 budget, with approximately 10 engagements for each dollar spent.</li></ul><p>The performance is clear: strategy 1 emerges as the winner based on the estimated average ROI. This represents the total engagement Alex receives per dollar spent on the influencers’ posts. Although the estimation has a wide range, suggesting a higher fluctuation in the estimated ROI value, Alex is okay with that! It provides him with a starting point.</p><p>Alright, Alex, let’s put it to the test. Go with the first strategy! Now all Alex needs to do is reach out to the influencers in those categories and start the collaborations. It’s important to keep in mind that the first few campaigns will always be experiments. This means Alex will have to diligently track the numbers and analyze the outcomes.</p><p>Good luck, Alex!</p><p>🔗 The calculator Alex uses is here (Download link): <a href="https://drive.google.com/file/d/1BpMpS16VlBsVhEn9W5AqF4YPneXeZc-s/view?usp=sharing">https://drive.google.com/file/d/1BpMpS16VlBsVhEn9W5AqF4YPneXeZc-s/view?usp=sharing</a></p><p>🔗 <a href="https://colab.research.google.com/drive/1S1A3i1x1Zv17zmhNYPiHBtN7AxzWRuqf?usp=sharing">You can also see the Colab version</a> right in your browser too, if you want.</p><p>👋 <strong>Additional help: </strong>if you are looking to adjust this calculator to your specific situation, <a href="https://calendly.com/peiran-yu/15min?back=1&amp;month=2023-09">don’t hold back! Simply reach out here for a quick chat</a><strong>.</strong> I’m also curious how it’d go for you.</p><h3>Some notes</h3><p><strong>Regarding how it’s used:</strong></p><p>This calculator is a simplified version that can be adjusted, adapted, or expanded to accommodate a wider range of scenarios. Without adjustments, it can be directly applied to the following scenario:</p><ul><li><strong>Known influencer categories or selected influencer individuals:</strong> You are considering influencers with varying levels of following and plan to purchase multiple posts from different levels. Alternatively, you have already selected a set of influencers and are narrowing down your choices among them. For instance, Tim, Kara, and Liane have approached you, offering their analytics deck and different price points. Now, you are deciding how many posts to purchase from each of these three influencers.</li><li><strong>Known budget:</strong> You have a budget that allows you to purchase posts from multiple influencers, and you have a good idea of the amount.</li><li><strong>Standard success metric:</strong> You define your return on investment (ROI) for all influencers based on one standard metric, such as engagement over cost. It is important to have a standardized definition of engagement, such as the total counts of likes, comments, shares, or just the total counts of shares if you want to focus on raising awareness.</li></ul><p><strong>Regarding its limitations:</strong></p><p>This calculator makes certain assumptions, for example, assuming that engagements are normally distributed. However, in real life, these assumptions may not always hold true. Additionally, the demographic match is assumed to be a static multiplier for each influencer, whereas in reality, this factor may have variability.</p><p>However, the accuracy of the calculator can improve if you learn from experiments and continuously collect more data. The quality of input data is just as important as the accuracy of the model within the calculator. For instance, if after collecting historical data on influencers, you have data indicating that shares per post don’t adhere to a normal distribution but more to a Poisson one, you can tweak the script to match.</p><p>Lastly, this calculator serves as a valuable reference, rather than a definitive answer, for your strategy. After it presents the strategy recommendations, users should leverage their industry expertise and consider all factors holistically to make the final decision.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=33d303af4bc0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[[Infographic] 290 Influential Gen AI Tools Advancing into 2024 as Category Leaders]]></title>
            <link>https://medium.com/@peiran.yuu/infographic-290-influential-gen-ai-tools-advancing-into-2024-as-category-leaders-f68b80d07a6a?source=rss-2e3799f0b15e------2</link>
            <guid isPermaLink="false">https://medium.com/p/f68b80d07a6a</guid>
            <category><![CDATA[gen-ai-tools]]></category>
            <category><![CDATA[growth-tools]]></category>
            <category><![CDATA[gen-ai-for-growth]]></category>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[gen-ai-for-startups]]></category>
            <dc:creator><![CDATA[Pei-Ran Yu]]></dc:creator>
            <pubDate>Wed, 06 Mar 2024 07:35:08 GMT</pubDate>
            <atom:updated>2024-06-07T09:06:03.984Z</atom:updated>
            <content:encoded><![CDATA[<p>(*Originally published on Jan 2024) As 2023 drew to a close, the AI arena has gotten more competitive than ever. Among the myriad of AI projects — over 8000 as of September 2023 — a select group of about 300 tools have emerged as the leading contenders.</p><h3>290 Influential Tools on The Infographic</h3><p>Based on the rankings from <a href="https://airankings.co/">airankings</a>, and to make it easier to analyze how the tools stand, I’ve compiled an infographic that details this ranking list ⬇️.</p><p>(It’s a pretty big file, please give it a few seconds to load)</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fview.genial.ly%2F65e81a7dc0aadd0014652487&amp;display_name=Genially&amp;url=https%3A%2F%2Fview.genial.ly%2F65e81a7dc0aadd0014652487%2Finteractive-content-medium-ai-tools-ranking-for-2024&amp;image=https%3A%2F%2Fthumbnails.genial.ly%2F658ce54e4eebb1001f8547dc%2Fscreenshots%2Fd3110651-f237-4d1d-ab5e-b2ff238084cc.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=genial" width="1200" height="675" frameborder="0" scrolling="no"><a href="https://medium.com/media/913be6459e70ba0263ac9f9222861099/href">https://medium.com/media/913be6459e70ba0263ac9f9222861099/href</a></iframe><p>In exploring the ranking list and the infographics, several captivating insights were uncovered.</p><h3>Overall</h3><ol><li>Top 3 tools in popularity: <a href="https://chat.openai.com/">ChatGPT</a> is the №1 tool among all, <a href="https://deepl.com/">DeepL</a> — a Translation tool at №2, and <a href="https://quillbot.com/">Quillbot</a> in Copywriting at №3;</li><li>Taking the top spot in the copywriting AI race isn’t <a href="http://copy.ai/">Copy.ai</a> or <a href="https://jasper.ai/">Jasper</a>, but <a href="https://quillbot.com/">Quillbot</a>, with <a href="https://wordtune.com/">Wordtune</a> following closely;</li><li>Contrary to expectations, <a href="https://pixlr.com/">Pixlr</a>, not Midjourney or Dall-E, takes the lead in image generation popularity;</li><li>The most populated categories in our list are <em>Image Generation, Image Editing, and Productivity. </em>The shakeout in 2024 is something to watch for for these categories. See here for the top 10 categories.</li><li>Top categories by average ranking: <em>Audio editing</em>, <em>Coding</em>, and <em>Photo editing</em>.</li><li>Notice many popular tools, like<a href="https://chatgptwriter.ai/"> ChatGPT Writer</a> (email tool) and <a href="https://pixlr.com/">Pixlr</a> (image generation), are all available for free.</li></ol><h3>Growth Related</h3><p>From a Growth perspective:</p><p>★ 1. Numerous tools now enable AI ChatBot creation, without coding and some without a credit card. The standout in this category is <a href="http://landbot.io/">landbot</a>. For a broader view, explore the “ChatBot” category in the infographic above.</p><blockquote>Startups lacking a customer support or sales chatbot should seriously consider adopting one in 2024.</blockquote><p>★ 2. Broader AI analytics tools face adoption challenges; only two have made it to the top list.</p><p>Highly ranked tools for growth teams to watch include:</p><ul><li><a href="https://forethought.ai/">Forethought</a>: a generative AI company that provides customer service automation.</li><li><a href="https://unbounce.com/">Unbounce</a>: a website builder that helps businesses create and optimize landing pages for better conversions.</li><li><a href="https://getresponse.com/">GetResponse</a>: an all-in-one online marketing platform for businesses to create, automate and analyze campaigns.</li><li><a href="https://framer.com/">Framer AI</a>: a versatile tool that builds marketing websites, landing pages, and online campaigns.</li><li><a href="https://frase.io/">Frase</a>: a content writing software that helps create SEO-focused content.</li><li><a href="https://landbot.io/">LandBot</a>: helps automate conversations for marketing, sales and customer service.</li></ul><h3>🎁 The Download Link</h3><p>You can download the interactive infographic PDF with 🔗 <a href="https://drive.google.com/file/d/1sCZIo3u_DIVJCbfE7x6YU4DI0EkOwaTl/view?usp=sharing">this link</a>.</p><p>Get ready! 2024 is going to be all about adoption. Some say it’s going to be the year of “Chief AI Officers”, driving the AI adoption wave.</p><p>Exciting times ahead! 🎉 See you in the new year!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f68b80d07a6a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[6 Ways Gen AI Tools Will Rapidly Unleash Data Accessibility]]></title>
            <link>https://medium.com/@peiran.yuu/6-ways-gen-ai-tools-will-rapidly-unleash-data-accessibility-f6e16e7ec6c3?source=rss-2e3799f0b15e------2</link>
            <guid isPermaLink="false">https://medium.com/p/f6e16e7ec6c3</guid>
            <category><![CDATA[growth-analytics]]></category>
            <category><![CDATA[startup-growth]]></category>
            <category><![CDATA[data-analytics]]></category>
            <category><![CDATA[growth]]></category>
            <category><![CDATA[gen-ai-tools]]></category>
            <dc:creator><![CDATA[Pei-Ran Yu]]></dc:creator>
            <pubDate>Wed, 06 Mar 2024 05:03:36 GMT</pubDate>
            <atom:updated>2024-03-06T05:08:17.030Z</atom:updated>
            <content:encoded><![CDATA[<h3>Common Dilemmas in Data-Driven Organizations</h3><p>If you’re in marketing, you know the frustration of needing quick data insights. You’ve probably found yourself waiting for the analytics team to deliver, and during that wait, you felt the urge to just learn SQL yourself to get the answers faster? Or maybe as a growth lead, you understand that qualitative data is invaluable for your strategy. But without a team of experienced data scientists to help you analyze customer feedback, you’ve likely wished you could decipher all that text data on your own, haven’t you?</p><p>In the top <a href="https://www.synq.io/blog/data-team-size-at-100-scaleups?utm_source=partner&amp;utm_medium=routz.tech&amp;utm_campaign=october">100 U.S. and European scale-ups surveyed</a>, only 1–5% of the workforce is part of the data team. Let’s be optimistic and say it’s 5%. That leaves 95% of the staff who could benefit from data to make better decisions. Simple math shows that each data team member would have to serve around 20 people. With such a wide range of requests and the need for both speed and accuracy, it’s no wonder that many questions go unanswered. Data teams are stretched to the limit, leaving much of the organization to operate without the benefit of data insights, unless they’re part of a culture that prioritizes experimentation.</p><p>Back in 2011, analytics became the buzzword as companies began migrating their data to the cloud. Everyone wanted to be data-driven, but most fell short. The barriers? Limited access to data, lack of strong data teams or data skills, and the absence of a well-organized data structure within the company.</p><p>2023 comes, and the new generative AI tools are set to transform how businesses make decisions with data.</p><h3>Gen AI Tools Coming to The Last Mile of Data Analytics</h3><p>First, what’s the “last mile” in data analytics all about? Imagine data going from the moment it’s generated and collected to when it’s processed, stored, and finally analyzed and visualized. The “last mile” is that crucial moment when the data is actually used and analyzed. (please note: don’t think Gen AI is just about changing the Last Mile. It’s gonna shake up the whole data game, from collecting to storing and more. But hey, if you’re in growth, let’s focus on how it’s changing that Last Mile for now.)</p><p>If you’re in growth, you know that even a little SQL knowledge can go a long way. If SQL isn’t your thing, you could find yourself stuck in the slow lane.</p><p>But here comes a game-changer: analytics tools supercharged by large language models.</p><p>These advanced tools are poised to revolutionize both data accessibility and speed to insights, making it quicker and easier for everyone to get the information they need.</p><p>Let’s check out the six cool tools we’ve found so far. (If you’ve come across something awesome that we missed, feel free to give us a heads-up!)</p><h3>№1 - Natural Languages to insights, SQL</h3><p><strong>Representing tool: </strong>Domo AI</p><p><strong>Tool announced: </strong><a href="https://www.domo.com/news/press/domo-announces-domo-ai-to-revolutionize-the-way-businesses-manage-and-deploy-artificial-intelligence?utm_source=partner&amp;utm_medium=routz.tech&amp;utm_campaign=october">2023/08</a></p><p><strong>What happens: </strong>non-technical business users can interact with data using natural language text to ask questions. This capability will provide: Text-to-SQL — generates SQL queries from natural language text. Text generation — automates writing processes, generating reports with less manual effort.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/329/1*rPqBq2L3rvfcWuxgrbn31g.jpeg" /></figure><p><strong>Implications to the growth team:</strong></p><p>Domo AI aims to solve the data dilemmas we talked about earlier by making data more accessible to more people in organizations. Think of it like chatting with your PDFs, but now you’re chatting with your databases.</p><p>One of the big wins here is how much easier it is to get your hands on <em>crucial growth metrics</em> and growth accounting data like MAUs, retention rates, and activation rates. What’s really cool? SQL is usually all it takes to dig up these numbers from your relational databases. With tools like Domo, everyone on the growth team — be it marketers, product managers, early-stage founders, or even the analysts themselves — can get to those key insights in no time flat.</p><p>Another significant advantage is how these tools <em>streamline the problem-solving process</em>. When juggling multiple tasks, marketers may notice a drop in acquisition rates, or product managers might see a dip in activation rates. Manually running various SQL queries to uncover the root cause can be intricate and lengthy. However, using a tool like Domo significantly speeds up this essential diagnostic process. There’s no need for a team-wide scramble or to wait for a data analyst. The issue is diagnosed and resolved on the spot.</p><p>By the way, while Domo has the backing of being an established company, it’s not the only tool available for translating natural language to SQL or insights. There are other tools like AI2SQL.io, which supports multiple languages — English, Spanish, German, French. Other options include -</p><ul><li><a href="https://aiquery.co/">https://aiquery.co</a></li><li><a href="https://outerbase.com/">https://outerbase.com</a></li><li><a href="https://aihelperbot.com/">https://aihelperbot.com/</a></li><li><a href="https://dbsensei.com/">https://dbsensei.com/</a></li></ul><h3>№2 - Natural Languages to statistical models, in Python &amp; R</h3><p><strong>Representing tool: </strong>Defog AI</p><p>Is a YC backed startup that turns natural language into SQL, Python and R by deploying fine-tuned LLMs; and helps speed up complex analysis by automating trial-and-error in statistical analyses.</p><p><strong>Tool announced: </strong><a href="https://www.linkedin.com/posts/defog-ai_excited-to-launch-defog-agents-today-these-activity-7115755670783590400-Bjhx?utm_source=share&amp;utm_medium=member_desktop">2023/10</a></p><p><strong>What happens:</strong> Data analysts and scientists’ roles get augmented by this tool, which auto-generates SQL and streamlines the trial-and-error process in Python &amp; R. Just put your analysis request in the interface, and you’ll get a preliminary model in Python &amp; R.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*teWoH_bFOgSaj_qo.gif" /></figure><p><strong>Full Video demo: </strong><a href="https://www.linkedin.com/posts/defog-ai_excited-to-launch-defog-agents-today-these-activity-7115755670783590400-Bjhx">see here</a></p><p><strong>Official site: </strong><a href="https://defog.ai/blog/agents/?utm_source=partner&amp;utm_medium=routz.tech&amp;utm_campaign=october">see here</a></p><p><strong>Implications to the growth team:</strong></p><p>Younger teams with just a few data scientists, who usually focus only on high-priority projects, can benefit in several ways:</p><ol><li>Boost the number of advanced analyses, taking on more predictive analysis like churn analysis, customer segmentation, and diagnostic studies using regression models.</li><li>Conduct statistical analyses for tricky scenarios that were previously low-priority or overlooked, like assessing the statistical significance and real impact of offline campaigns.</li><li>Empower data scientists to think creatively, such as using linear programming to make the most of paid ad budgets.</li></ol><h3>№3 - SQL Functions for Unstructured Data Analysis</h3><p><strong>Representing tool: </strong>DataBricks AI functions</p><p><strong>Tool announced: </strong><a href="https://www.databricks.com/blog/2023/04/18/introducing-ai-functions-integrating-large-language-models-databricks-sql.html?utm_source=partner&amp;utm_medium=routz.tech&amp;utm_campaign=october%5C">2023/04</a></p><p><strong>What happens: </strong>SQL Analysts without unstructured analysis skills can now analyze complex unstructured data, in their familiar SQL environment. For instance, you can use a simple SQL function to send a request to OpenAI’s GPT models. This can turn an audio file, like the mp3s you see in the screenshot below, into text. Another AI function can then summarize that text, and you get the summary right in your SQL table.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*KhiUY-RCdJoeu3Vi.png" /></figure><p><strong>Full Video demo: </strong><a href="https://www.youtube.com/watch?v=ndL3TlWlA1o&amp;t=14s">see here</a></p><p><strong>Official site: </strong><a href="https://www.databricks.com/blog/2023/04/18/introducing-ai-functions-integrating-large-language-models-databricks-sql.html?utm_source=partner&amp;utm_medium=routz.tech&amp;utm_campaign=october">see here</a></p><p><strong>Implications to the growth team:</strong></p><p>As your SQL analysts get the hang of analyzing unstructured data, two things become key:</p><ul><li>Get strategic: Start asking the right questions, especially since it’s now easier to get answers.</li><li>Source more unstructured data: Now that you can analyze it, where do you find it? Consider tapping into less traditional data sources like community growth metrics or user feedback, whether direct or from social media.</li></ul><p>DataBricks is really leading the pack! While they offer a lot, another standout feature is their “DataBricks Assistant.”</p><h3>№4 - Natural Languages to SQL &amp; Python code</h3><p><strong>Representing tool: </strong>DataBricks Assistant</p><p><strong>Tool announced: </strong><a href="https://www.databricks.com/blog/introducing-databricks-assistant?utm_source=partner&amp;utm_medium=routz.tech&amp;utm_campaign=october">2023/07</a></p><p><strong>What happens:</strong> data analysts and data scientists can collaborate in code-writing with the DataBricks assistant, much like a co-pilot experience.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*iMWe7Dc_1sqBEbXo.gif" /></figure><p><strong>Full Video demo: </strong><a href="https://www.youtube.com/watch?v=g9REIGKkAVo">see here</a></p><p><strong>Official site: </strong><a href="https://www.databricks.com/blog/introducing-databricks-assistant?utm_source=partner&amp;utm_medium=routz.tech&amp;utm_campaign=october">see here</a></p><p><strong>Implications to the growth team:</strong></p><p>Augmented — picture superhuman data analysts and data scientists — translates to a higher volume of data requests fulfilled, leading to more data-driven organizations! Data scientists could also tackle the more challenging tasks.</p><h3>№5 - Natural Languages to SQL, via Python</h3><p><strong>Representing tool: </strong>LangChain</p><p><strong>Tool announced: </strong><a href="https://en.wikipedia.org/wiki/LangChain">2023/04</a></p><p><strong>What happens:</strong> Python coders can create easy-to-use data apps instead of just making dashboards. These apps let other folks in the business, who might not be tech-savvy, pull data from databases and get insights themselves.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/938/0*t-GqOFAamQFRsbHZ.png" /></figure><p><strong>Full Video demo: </strong><a href="https://www.youtube.com/watch?v=Fz0WJWzfNPI">see here</a></p><p><strong>Official site: </strong><a href="https://python.langchain.com/docs/use_cases/qa_structured/sql">see here</a></p><p><strong>Implications to the growth team:</strong></p><p>Python becomes an imperative for small growth teams. This enables the scenarios described earlier, as well as the development of MVPs. Look for team members experienced with Python.</p><p>Now if that doesn’t already grab your attention, check this out:</p><h3>№6 - Natural Languages to charts, with CSV files</h3><p><strong>Representing tool: </strong>DataSquirrel</p><p><strong>Tool announced: </strong><a href="https://www.linkedin.com/posts/datasquirrel-ai_datasquirrel-ai-your-fastest-way-from-csv-activity-7061970628865650688-cP4h">2023/05</a></p><p><strong>What happens: </strong>your data cleaning and analysis time drops from hours to just minutes. Plus, you get all sorts of charts almost instantly. All you need to do is upload a CSV file.</p><p><strong>Full Video demo: </strong><a href="https://www.youtube.com/watch?v=ohbuGbUvJNU">see here</a></p><p><strong>Official site: </strong><a href="https://datasquirrel.ai/?utm_source=partner&amp;utm_medium=routz.tech&amp;utm_campaign=october">see here</a></p><p>Not everyone is well-versed in SQL or Python, and the same goes for Pivot Tables. So, what’s the big deal about how you present your data? Well, the format of your charts can seriously influence the insights you gain, making diverse data visualization options important.</p><p>Why?</p><p>Consider a straightforward Excel table for a SaaS company with the following columns:</p><ul><li>User ID</li><li>Joining month</li><li>Whether or not they converted</li><li>Converted month</li></ul><p>What questions come to mind based on this data? What insights could you potentially extract? You might first think of calculating the monthly conversion rate. But wait — how exactly do you define that? Is a January conversion a user who joined and converted in the same month? What if your product has a lengthy sales cycle? Maybe you’d like to consider users who joined in January but converted in a later month. Do you group the data by “joining month” and then compare total users to those who converted? Or could you visualize it in a more revealing way?</p><p>The manner in which you present this data can yield vastly different insights.</p><p>(Excuse the doodle-like quality of these hand-drawn charts; they’re just to quickly illustrate the point.)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*-EHGW1ssWbDDtpYr.png" /></figure><p>Take a look at the chart on the left. The lines represent different cohorts — users who joined in specific months. The X-axis shows the total conversion rate by each month. This simple chart tells you a lot:</p><ul><li>What’s the average conversion rate like?</li><li>Is January’s conversion higher than March’s?</li><li>How quickly do clients usually convert? If the line has a gentle slope, conversions are slow. But if there’s a steep rise at Month 3 (M3), you know it typically takes people three months to convert.</li></ul><p>Once you’ve got a handle on the big picture, you can dig deeper with a second chart. Say you’re wondering why the January cohort stopped growing in Month 3. Was there a hiccup in the sales funnel? Chart 2 can give you those answers.</p><p>This is a prime example of why a simple single chart won’t be enough, even with simple spreadsheets, the right data presentation can be a game-changer, offering a ton of insights in just one chart. The catch? Not everyone knows how to create these charts, and that’s a challenge in itself. But here’s the good news: upcoming generative AI analytics tools promise to make this task a whole lot easier, leading to a substantial increase in insights, with just those simple spreadsheets.</p><p><strong>Implications to the growth team:</strong></p><p>Frontline decision-makers, like product managers, need to really up their game in business smarts and growth know-how. For instance, the ability to quickly understand things like cohort analysis and heat maps becomes crucial.</p><p>If most teams jump on the tool bandwagon, growth teams face a more challenging question: “What’s the next step?” Growth teams are all about executing a solid strategy, and strategy revolves around gaining a competitive edge.</p><blockquote>So, if everyone has these tools, how can you help your company stand out? What new areas should you be keeping an eye on?</blockquote><h3>Human Growth Team: 3 Ways to Go the Extra Mile</h3><p>When it comes to AI, especially generative AI, it produces unanimity. In other words, it provides fast and standardized answers for everyone.</p><p>For startups, this means you can get your activation insights quickly, but so can your competitors. When everyone relies on unanimous solutions, it won’t set you apart. So, what’s the key to standing out? It’s not just about optimizing the last mile; you have to go the extra mile, or even extend it before the journey begins. Here’s what that entails:</p><h4>1. Hire people that are truly curious.</h4><p>First and foremost, hire individuals who possess genuine curiosity. Curiosity works like magic for those in growth roles. It’s the unseen force that drives people to ask additional questions beyond what a typical gen AI tool might come up with. These unique and valuable questions can lead to unique value propositions that others in your industry might overlook.</p><p>Curiosity also motivates your team to not simply accept data at face value but to pursue the more challenging data — the qualitative kind obtained through conversations with your customers. This often turns out to be a goldmine because it can be the source of unique insights that lead to distinctive actions your competitors aren’t taking.</p><h4>2. Establish a robust process for your even more rapid experimentation pipeline.</h4><p>As data becomes more decentralized, companies are poised for greater innovation. This is because they’re placing data and valuable insights directly into the hands of those dealing with the relevant problems. This concept is also discussed in the podcast episode 🎙️ titled “<a href="https://open.spotify.com/episode/7GdhV9SI9uY7Tkqp0KNyFE">Why Online Retailer Zelando Was First to Embrace the Data Mesh</a>” on May 3rd, around the 27-minute mark.</p><p>A flawed process might look like this: your product team has significantly different key performance indicators (KPIs) compared to your growth team, leading to competition rather than collaboration. There’s no comprehensive growth performance dashboard to track how everything aligns with your primary goal. Your experimentation pipeline isn’t even set up, and experiments happen haphazardly, without a connection to your primary goal. Additionally, your frontline colleagues rarely communicate with your data or growth teams.</p><p>When your process lacks structure, empowered team members with access to more data insights may find themselves facing more challenges than enjoying the benefits.</p><h4>3. Strategize data collection and ensure data observability in your organization.</h4><p>In a world where you can’t collect every piece of data, it’s essential to focus on gathering the truly important data points. This might involve creating well-designed surveys sent to specific customers regularly, considering the purchase of external data sources that your competitors haven’t explored, or even collecting more unstructured data from your online community groups.</p><p>As insights reach decision-makers at a faster pace and come from various data sources, the quality of data becomes a critical concern. Just imagine the potential disaster if frontline workers were making decisions based on incorrect numbers. With AI making data accessible to more individuals, the scale of such a disaster could escalate rapidly.</p><p>(By the way, the founders of <a href="https://www.montecarlodata.com/?utm_source=partner&amp;utm_medium=routz.tech&amp;utm_campaign=october">Monte Carlo</a>, Barr Moses and <a href="https://www.linkedin.com/in/lgavish/">Lior Gavish</a>, are the ones who coined the term ‘data observability.’ If you’re keen on delving deeper into the realm of data observability and data quality, you won’t want to overlook this tool or their book “<a href="https://www.amazon.com/Data-Quality-Fundamentals-Practitioners-Trustworthy/dp/1098112040">Data Quality Fundamentals</a>”.)</p><p>As you embrace these three suggestions, you’ll find yourself well on your way to going the extra mile!</p><blockquote>That extra mile is what truly what’s going to separate you apart and lead you to a winning position.</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f6e16e7ec6c3" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Channeling Chaos into Effective Growth Experiment Pipelines]]></title>
            <link>https://medium.com/@peiran.yuu/channeling-chaos-into-effective-growth-experiment-pipelines-2f35639ba22e?source=rss-2e3799f0b15e------2</link>
            <guid isPermaLink="false">https://medium.com/p/2f35639ba22e</guid>
            <category><![CDATA[growth-strategy]]></category>
            <category><![CDATA[growth-analytics]]></category>
            <category><![CDATA[growth-hacking]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[Pei-Ran Yu]]></dc:creator>
            <pubDate>Tue, 05 Mar 2024 07:34:08 GMT</pubDate>
            <atom:updated>2024-03-06T07:39:44.097Z</atom:updated>
            <content:encoded><![CDATA[<p>Ever found yourself wonder -</p><blockquote>“With so many possible growth tactics, which one should I focus on, right now?”</blockquote><p>If this question has you stumped, you’re not alone. Prioritization is a hurdle many fast-moving startup founders and growth leaders face. You might be debating whether to invest in paid ads, enhance organic traffic, or maybe bring in someone to drive community growth. Choosing the right CRM tool or fine-tuning pricing are also critical decisions. Young companies are especially rich in ideas but often limited by team size and resources. It’s quite like a contestant on “MasterChef” quickly deciding what to cook from an array of ingredients, all while under time constraints — definitely no easy feast!</p><p>To find the right focus, start by <strong><em>doing some simple math</em></strong>. Analyze the data to pinpoint the areas with the most significant impact. Once you’ve identified these areas, <strong><em>conduct targeted experiments</em></strong> to validate your ideas specifically for them.</p><p>Sounds great to you but unsure about how to go about it?</p><p>Seedstars, a venture capital platform and accelerator that touts its growth program as its superpower, advocates an innovative approach known as the “Control Tower”. This article will take your through the two-step process of the Control Tower and how it can help startup founders like you in identifying and systematically addressing growth opportunities with precision.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*O7CGPyBVa7FBvyva006m0A.png" /></figure><p>Control Tower is set up within a <strong><em>Google Sheet</em></strong>. Think of it as part simplified BI dashboard and part simplified experimentation roadmap. It’s grounded in the well-known pirate funnel structure, <strong><em>AARRR</em></strong>, to provide a holistic view of growth performance. The system includes three key elements — the AARRR Funnel, Sprints and Experiments — creating a structured approach to test ideas with <strong><em>clear prioritization</em></strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*51AK9rhNJBYSv6caGYOd2g.png" /></figure><p>Let’s explore each component in detail to give you a clear understanding.</p><h3>Step 1 — Identify opportunities using simple math</h3><p>And the -</p><h4>The AARRR funnel</h4><p>Kick things off with this essential overview of your growth performance. The AARRR Funnel is exactly what it sounds like: a snapshot of your key metrics in Acquisition, Activation, Retention, Revenue, and Referral. In this step, you’d want to stick to a maximum of three main metrics per category to keep it neat and focused. This approach helps you get a quick, clear view of where the big opportunities are lying.</p><p>This is what the “AARRR Funnel” tab looks like on the Control Tower Google Sheet -</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xIVOun9pz4B9BR7gQbz4Pg.png" /></figure><p>Right at the top, you’ll see your North Star Metric. Understanding your North Star Metric is crucial, as it aligns all your growth initiatives and team efforts, giving you clear direction and focus on what matters most.</p><p>Each metric starts with an “User Action” — that key action users take, crucial for the success of each category. Then, the “Metric” gives the names of metrics related to these actions. We also include the “Data Source” for each metric, as you might need to investigate further if numbers seem off. Decide whether to display your metric as an absolute number, like “Organic Website Visits”, or as a rate, such as “monthly retention rate”, and record these in separate columns. These are usually presented as Monthly Averages for simplicity. This is particularly effective if your opportunity size is stable, where a “quarterly monthly average” might be used. But if your situation frequently changes and you’re constantly trying new ideas — which is highly recommended — a “Last Month” view might be more beneficial than a “Monthly Average”. We’ll delve deeper into this soon, but let’s hold that thought for now.</p><p>Finally, we get to the crucial part — determining “Priority”. Once you’ve defined your metrics and filled in the numbers, let this sheet guide your focus. Choose up to three top growth priorities, or just one if you really want to zero in on a single area. But how do you pick these priorities?</p><p>Look at the “Monthly Average” columns to spot the top three metrics that, if improved, would most positively affect your North Star Metric. These are your key growth opportunities for now. You might need to play “defense” — zeroing in on metrics that have shown a significant drop from month to month, signaling potential areas for improvement. Alternatively, consider “offense” — setting ambitious targets for specific metrics and focusing on those that are the furthest from these goals as your main areas for growth.</p><p>Picture the five metric categories in our spreadsheet as a set of five cards. Each card vividly represents a category, from Acquisition to Referral. As you analyze them, it’s like weighing the value of each card.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/452/1*luWQmm2s2TTX3Yxj08Nxow.jpeg" /><figcaption>When you pick the two priority metrics, it’s as if you’re choosing the two best cards to play, bringing those strategies into action.</figcaption></figure><p>For young startups in the pre-seed to seed stages, it’s common to focus on Awareness rather than immediate revenue, thanks to their financial runway. In these situations, an AAARR structure — Awareness, Acquisition, Activation, Retention, Referral — is a good fit. Alternatively, an AAARRR structure can be used to include Revenue. The key is to adapt your focus as your growth strategy evolves.</p><blockquote>If this concept seems abstract, no need to worry! You’re covered with an example at the end to enhance your understanding. And remember, <a href="https://www.linkedin.com/in/yupeiran/">don’t hesitate to reach out to me on Linkedin if you have any questions.</a></blockquote><h4>What if Clubhouse adopted this Control Tower view?</h4><p>Remember when Clubhouse took the world by storm in 2020? It was the hottest app, attracting celebrities and creating a buzz. If you were lucky enough to receive an invitation from a friend, you owed them a big thank you. The hype was incredible, and Android users felt like they were missing out. Fast forward three years, X (formerly Twitter) launched Twitter Spaces, Discord introduced Discord Stage Channels, and even LinkedIn joined the trend with LinkedIn Audio Events. Despite the advantage of a massive user base for these bigger platforms, Clubhouse still had a fighting chance. They had a solid 8 months to focus on retention before Twitter Spaces arrived in November 2020.</p><p>Did they not realize that their retention was the issue? Perhaps they did. Check out this alarming retention cohort chart for Clubhouse found on the Internet -</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ObQUxObgwKzN1JhnXMxbqw.png" /></figure><p>Why then didn’t they prioritize retention? Were they overwhelmed with ideas and struggled to prioritize? Perhaps they were too focused on achieving high activation rates in order to onboard new users coming in large volumes.</p><p>Imagine if they had a Control Tower — it would be screaming “Retention” and “Revenue” right at them! In this fictional and simplified Control Tower for Clubhouse below, the problematic areas are highlighted in red, showing the not-so-great Retention and Revenue numbers. On the bright side, the growth areas, where things are going well, are highlighted in green, representing acquisition/referral and potentially activation. It makes sense that, with this setup, you’d expect them to prioritize “Retention” as a top priority.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/820/1*3WY7wmYCD99ffRzN4FNr_g.png" /></figure><p>Great, now you have seen the first step, the rest that follows will only seem natural to you — now that you narrowed down to a few important growth areas, it’s time to think of some ideas to experiment with to increase those numbers. Experiments should also be run in a structured way, that’ll be covered next.</p><h3>Step 2 — Test ideas systematically with</h3><h4>Sprints and Experiments</h4><p>Each sprint corresponds to a crucial metric you’ve picked from the “Priorities” in the AARRR Funnel step. Imagine your retention is at a concerning 30%, well below the industry standard of 80%, but your other metrics look good. In this scenario, your aim would be “boosting monthly retention to 80%”, and this goal becomes your “Sprint Objective” in this step.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rIWE3QsGobD_7WCduZC8YA.png" /></figure><p>Experiments sit a level below “Sprints”, each experiment corresponding to either a hypothesis or idea focused on a growth opportunity or challenge. So, if you’re honing in on the sprint for the retention metric, you’d generate ideas aimed at boosting that particular metric. Let’s take Clubhouse again as an example. A hypothesis for their low retention might be the ephemeral nature of their audio content. Users who return have limited activities, maybe just profile browsing. This leads to ideas like creating robust community tools for ongoing conversations, adding gamification elements like badges, or allowing content creators to post summaries of past events, engaging users further. Each of these Ideas then corresponds to a line in the “Experiments” step, or a row in this part of the Control Tower, all focused on the same Sprint Objective: “Improve Retention Metric to 80%”. And so,</p><p><strong><em>Phase 1: Brainstorm</em></strong></p><p>Looks like this -</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/940/1*r6q-SsRXFV9kNCwobfrMqQ.png" /></figure><p>Next, take the prioritization a step further at the idea level using the ICE method. Give each idea a prioritization score, which then helps you set the appropriate “Priority” level for each. With this approach,</p><p><strong><em>Phase 2: (Further) Prioritize</em></strong></p><p>Shapes up like this -</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/621/1*IcYFF0eeziCxTAhgEnTxQQ.png" /></figure><p>This is based on the assumption you’re up to speed with the ICE methodology for prioritization. If that’s not the case, no worries — just give it a quick Google and check out a few guides.</p><p>Now that you’ve got your priorities lined up, it’s time to assign an ‘Owner’ and a date range for each experiment you’ve picked. Some experiments may take longer and need an early start, while others, quicker to complete, can slot in any time during the sprint’s schedule. Also, begin the “High” priority experiments earlier, and consider postponing or skipping the ‘Low’ priority ones unless you find some extra time. At the experiment level, set an expected number for the Sprint metric, aiming for all ideas to collectively hit your Sprint objective’s targeted increase. With this setup,</p><p><strong><em>Phase 3: Plan</em></strong></p><p>Comes together like this -</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/846/1*zoIHTcy7K1q4jOAVwx5B2w.png" /></figure><p>Finally, add the execution specifics, including a link to your experiment design document. Your design should cover several key components like the success metric, target audience, control versus experimental groups, and tracking methods. In the Control Tower, simply include a link to this document to keep things streamlined. Also, jot down a quick summary of Final Results, Learnings, and Next Steps, enabling future quick reviews of key insights. This approach forms</p><p><strong><em>Phase 4: Execute</em></strong></p><p>which is laid out like this -</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/963/1*gABCWRY1sbiJBBQW016wWA.png" /></figure><p>Since a lot of experiments typically don’t pan out, it’s smart to run a good number of them to make sure you hit your Sprint goals. The idea is to have a steady flow of ideas, all neatly prioritized. Find the ones that click, then work on systemizing and scaling them up for bigger and continuous impact.</p><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*O7CGPyBVa7FBvyva006m0A.png" /></figure><h4>An example and a template for startup founders</h4><p>To better demonstrate these concepts, a fictional company, “Big Bite”, was created to serve as an example.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/120/1*9WmD3_1a3xWANJXYC4wcEQ.png" /><figcaption>Imagine Big Bite as a food delivery app operating on a subscription model, currently active in Singapore, and gaining some initial success. They’re aiming to expand their brand to more global locations, starting with the South East Asia market.</figcaption></figure><p>Let’s assume they’re new in the market and, according to their financial projections, can concentrate on acquiring and activating customers without immediate concern for revenue for a few months.</p><p>Wondering what their Control Tower might look like? Take a peek at the AARRR component for this fictional -</p><p>🔗 <a href="https://docs.google.com/spreadsheets/d/1ix-xBEMOU4_s-NWI6L9JwEyCxMaAZ5KZ5Ckf2HhnfWA/edit?usp=sharing"><strong>Control Tower for Big Bite</strong></a></p><p>The Control Tower offers key benefits when you’re swamped with countless potential tactics. It keeps you focused and provides a big-picture view, showing how different growth levers interlink and highlighting gaps and opportunities. Importantly, it reminds you that growth tactics should be rooted in a data-driven growth strategy. While embracing creativity, any new idea should be assessed for its fit in the Control Tower framework and its impact on your North Star Metric and long-term success.</p><p>Ready to dive in? Here’s a -</p><p>🔗 <a href="https://docs.google.com/spreadsheets/d/1IORLDsVpWZRVknWsfFjA3apKsSAbA8LisMzb9wAZoi0/edit?usp=sharing"><strong>Control Tower Template</strong></a></p><p>crafted to get you up and running with the Control Tower.</p><p>If you want to learn more about how to prioritize your growth experiments, feel free to <a href="https://www.linkedin.com/in/yupeiran/">connect with me and message me on LinkedIn</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2f35639ba22e" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>