<?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 Abhishek Patnaik on Medium]]></title>
        <description><![CDATA[Stories by Abhishek Patnaik on Medium]]></description>
        <link>https://medium.com/@abhishekpatnaik77?source=rss-3b21676c7da2------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*_EqcV-gE3EpDTDx9x-GXcQ.png</url>
            <title>Stories by Abhishek Patnaik on Medium</title>
            <link>https://medium.com/@abhishekpatnaik77?source=rss-3b21676c7da2------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 22:46:33 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@abhishekpatnaik77/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[Are You Suffocating Your Users With Limits?]]></title>
            <link>https://medium.com/surge-startup/are-you-suffocating-your-users-with-limits-6e8fe9236e8d?source=rss-3b21676c7da2------2</link>
            <guid isPermaLink="false">https://medium.com/p/6e8fe9236e8d</guid>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[business]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[Abhishek Patnaik]]></dc:creator>
            <pubDate>Tue, 23 May 2023 09:41:07 GMT</pubDate>
            <atom:updated>2023-05-23T09:41:07.499Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/612/1*YzA71qzBbTzyKg0eD34ygg.jpeg" /></figure><p>Albert Einstein said, “Once we accept our limits, we go beyond them”. The same goes for the user who is using your product. They will go beyond your tool and start using some other tool if you limit your product too much. LOL!</p><p>Remember? Excess of everything is harmful.</p><p>In today’s blog, we keep things simple, we discuss how you can create limits according to pricing. Thereby helping you with a significant profit margin.</p><h3>What are Limits?</h3><p>Limits by definition mean imposing restrictions on the complete app depending upon different plans.</p><p>For example, your tool has 3 features namely X, Y and Z. You choose to restrict Y,Z only to premium customers. That’s what limits are.</p><h3>How does that Impact my product?</h3><p>It impact’s your product largely. As a Startup, you want to make sure that users are able to explore the solution to their problems. You cant just restrict the USP of your tool or make limits way too complex for users to digest it.</p><h3>Suffocating the users with Vague Limits</h3><p>Often CEO, of the company tends to make a pricing offer that starts suffocating the users and they quit using it. Your product might be a ground breaker but it will fail if users find it hard to use.</p><p>Let me give you an example from my previous experiences,</p><p>I was working as a product manager of a startup a few years ago. The CEO of the company asked us to give an estimate of the server costs so that an efficient pricing plan could be made.</p><p>The CEO got inspired by a game made for kids. And decided to add a token for every feature. Now think, how confusing would it be, there were a total of 12 different features inside the product. 12 different feature has 12 different ways of charging token.</p><p>Here’s an example of how the pricing page looked.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EgbM6YNus7FeF56lWOYsPg.png" /></figure><p>How did it impact the product?</p><p>Well every time, a feature was used customers would be concerned about losing credits. They have to always check the limits page, to confirm that the feature they were using, if that would end their credits.</p><p>This made the app usage quite difficult. Now imagine, google starts charging you based on the length of the search query. You would feel frustrated right !! You would switch to some other browsers like Bing or Brave.</p><p>The same implies here, the users still think that the product is difficult to use, because of the poor product design.</p><h3>How to FIX this ISSUE?</h3><p>Here are a few ways you can make sure that your limits are not killing your customer&#39;s interest.</p><ol><li>CEOs often take decisions based on their gut feeling, without considering the views of a product expert. It becomes super important to consult with the team and the manager on how the pricing should be.</li><li>Always respect the customer&#39;s opinion. After we launched the limit change, the user was blown by storm with outrage. The best solution was to remove the limiter with a new easy to understand pricing.</li><li>Add a room to upsell your product. You would always want your user to switch to a premium plan. Make sure that you give them “trial access”, so that they can explore the different possibilities of your tool.</li><li>Never allow freemium plans(until you have loads of money to spend). If you are offering things for free to your users, you are going to face difficulties after you embed pricing over those free plans. Users won&#39;t be ready to pay for it. For example, Phone-pay, a fintech startup, started charging 0.05%-0.2% of phone recharge after giving recharge for free to users. They created a huge customer base, and now they can charge easily. But this happens only when you have a lot of money to build that kind of valuation. If you don’t have money simply don’t offer freemium plans.</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6e8fe9236e8d" width="1" height="1" alt=""><hr><p><a href="https://medium.com/surge-startup/are-you-suffocating-your-users-with-limits-6e8fe9236e8d">Are You Suffocating Your Users With Limits?</a> was originally published in <a href="https://medium.com/surge-startup">Surge Startup</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AWS vs Google Cloud? What’s the best for ML Startup]]></title>
            <link>https://medium.com/surge-startup/aws-vs-google-cloud-whats-the-best-for-ml-startup-1085506d659f?source=rss-3b21676c7da2------2</link>
            <guid isPermaLink="false">https://medium.com/p/1085506d659f</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[aws]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[google]]></category>
            <dc:creator><![CDATA[Abhishek Patnaik]]></dc:creator>
            <pubDate>Wed, 10 May 2023 12:26:21 GMT</pubDate>
            <atom:updated>2023-05-10T12:26:21.226Z</atom:updated>
            <content:encoded><![CDATA[<h3>AWS vs Google Cloud? What’s the best for ML Deployments</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/563/1*h2TijLknySqlxreb4Zy61w.jpeg" /></figure><p>Wrong Decisions for the cloud might sometimes be a nightmare for your machine learning projects.</p><p>Machine learning has become an essential part of many businesses today, and with the increasing demand for data processing, cloud service providers have become increasingly important for organizations looking to leverage machine learning. Amazon Web Services (AWS) and Google Cloud are two of the most popular cloud service providers that offer machine learning solutions. However, deciding which one to choose can be a daunting task. In this blog, we will explore the strengths and weaknesses of both AWS and Google Cloud to help you make an informed decision.</p><p>But why do we need to know about this now? Well, I have been working and leading startups for the past 6 years. The most basic problem I see with startups is that they are confused about which cloud provider to use for their machine learning deployments.</p><p>Let’s get started by comparing them.</p><h3>Amazon Web Services (AWS):</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*Hx3yg42_2fQQqMbXXYlQ2g.jpeg" /><figcaption>AWS services</figcaption></figure><p>AWS is one of the largest cloud service providers and offers a wide range of machine learning tools, including Amazon SageMaker, a fully-managed service that enables you to build, train, and deploy machine learning models. AWS also offers a range of pre-built machine learning models, including image and text analysis, speech recognition, and natural language processing.</p><p><strong>Services:</strong></p><p>These are different services that AWS provides for Machine Learning.</p><ul><li>Amazon Lex brings Natural Language Processing toolkit and speech recognition possibilities, focusing on integrating <a href="https://www.projectpro.io/article/python-chatbot-project-learn-to-build-a-chatbot-from-scratch/429">Chatbot applications</a>.</li><li>Amazon Rekognition is a computer vision suite that renders the development and testing of face/object recognition models. It can easily perform complex CV tasks like object classification, scene surveillance, and facial analysis.</li><li>Amazon SageMaker is a full-fledged machine-learning platform that runs on EC2 instances and can develop traditional machine-learning implementations.</li></ul><p>Pros:</p><p>1. Broad Range of Tools: AWS has an extensive range of machine learning tools that cater to different requirements, including supervised and unsupervised learning algorithms. This makes it easy for developers and data scientists to choose the right tool for their specific use case.</p><p>2. Scalability: AWS provides on-demand resources, which makes it easy to scale up or down as needed. This ensures that your machine learning models can accommodate the varying demands of your business.</p><p>3. Integration: AWS integrates easily with other AWS services, which allows you to build and deploy complex machine learning workflows without worrying about compatibility issues.</p><p>Cons:</p><p>1. Complexity: AWS has a steep learning curve, which can be daunting for beginners who are not familiar with its services.</p><p>2. Cost: AWS pricing can be complex and difficult to understand. However, there are several cost-saving measures that you can implement to optimize your costs.</p><p>3. Data Transfer and Storage: Some users have reported difficulty with data transfer and storage in AWS.</p><h3>Google Cloud:</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*24ZUej4xhE2EZg4RHOPDcA.png" /><figcaption>Google Cloud Services</figcaption></figure><p>Google Cloud is a newer cloud service provider that offers a range of machine learning tools, including Google Cloud AI Platform, a fully-managed service that allows you to build, train, and deploy machine learning models.</p><p>Services:</p><ul><li>Google Cloud AutoML is a machine learning toolkit explicitly built for beginners in the field. It offers functionalities like data model upload, training, and testing through its web interface. AutoML integrates well with other Google Cloud services like cloud storage. It can perform all the complex machine learning problems like <a href="https://www.projectpro.io/project-use-case/face-recognition-python-project">Face Recognition</a>, etc.</li><li>Tensorflow: Tensorflow is an already renowned name in the machine learning community. TensorFlow is an open-source library for numerical computation and analysis. It is used widely in <a href="https://www.projectpro.io/article/deep-learning-algorithms/443">deep learning models</a> and packs many useful Machine Learning functions.</li><li>Vertex AI is an <a href="https://www.projectpro.io/article/mlops-projects-ideas/486">MLOps</a> platform that promotes experimentation through pre-trained APIs for natural language processing, image analysis, and computer vision.</li><li>Google Machine Learning Engine: It is the machine learning offering at scale from Google. Google ML engine can perform complicated Machine Learning tasks using <a href="https://www.projectpro.io/article/gpus-for-machine-learning/677">GPU</a> and Tensor Processing Unit while running externally trained models. With great efficacy, Google Machine Learning Engine automates resource provisioning, monitoring, model deploying, and hyperparameter tuning.</li></ul><p>Pros:</p><p>1. Cutting-Edge Tools: Google Cloud is known for its cutting-edge machine learning tools and technologies. Its tools are designed to provide fast and efficient processing for large datasets.</p><p>2. User-Friendly Interface: Google Cloud’s user interface is intuitive and user-friendly, making it easy for developers and data scientists to get started with machine learning.</p><p>3. Innovation: Google Cloud is known for its innovative machine learning tools and technologies, making it an ideal choice for organizations that need access to the latest advancements in machine learning.</p><p>Cons:</p><p>1. Data Transfer and Storage: Some users have reported difficulty with data transfer and storage in Google Cloud.</p><p>2. Cost: Google Cloud pricing can be complex, and users may find it difficult to understand how to optimize their costs.</p><p>3. Smaller User Community: Google Cloud has a smaller user community compared to AWS, which may make it more difficult to find support and resources.</p><p>Conclusion:</p><p>Both AWS and Google Cloud offer powerful machine-learning tools and technologies. The right choice depends on your specific needs and requirements. If you are already using AWS services and are familiar with their pricing structure and tools, AWS may be the better choice for you. However, if you are looking for cutting-edge machine learning tools and technologies or prefer Google Cloud’s user interface and pricing structure, then Google Cloud may be the better choice for you.</p><p>So Yes, Ultimately, both providers are capable of delivering powerful machine-learning solutions. It’s important to evaluate your needs and do your research to determine which provider is the right fit for your organization. Whatever your decision, make sure to explore the provider’s ecosystem, available support, and pricing models to ensure that you get the most out of your machine-learning solution.</p><p>Want to get in touch with me?? Feel free to DM me on LinkedIn:</p><p><a href="https://www.linkedin.com/in/abhishekpatnaik77/">https://www.linkedin.com/in/abhishekpatnaik77/</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1085506d659f" width="1" height="1" alt=""><hr><p><a href="https://medium.com/surge-startup/aws-vs-google-cloud-whats-the-best-for-ml-startup-1085506d659f">AWS vs Google Cloud? What’s the best for ML Startup</a> was originally published in <a href="https://medium.com/surge-startup">Surge Startup</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How does TB cause Startups to fail?]]></title>
            <link>https://medium.com/surge-startup/how-does-tb-cause-startups-to-fail-f933063d8e39?source=rss-3b21676c7da2------2</link>
            <guid isPermaLink="false">https://medium.com/p/f933063d8e39</guid>
            <category><![CDATA[product-development]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[saas]]></category>
            <dc:creator><![CDATA[Abhishek Patnaik]]></dc:creator>
            <pubDate>Fri, 05 May 2023 12:53:49 GMT</pubDate>
            <atom:updated>2023-05-05T12:56:24.663Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KuT7WJbEHBF4GuM0LwFVow.png" /><figcaption>Technical Debt</figcaption></figure><p>Around 90% of SaaS startups fail to make a dent in the market. The reasons are simple but we ignore them because of our skeptical behavior.</p><p>To be true, there are many reasons involved behind a startup’s failure. I have personally worked with more than 5 startups. During these years I have worked as a developer, Product Manager, Tech Lead, and Co-Founder. The reasons I state in this blog might be something different.</p><p>It’s cool to start building your own product, initially, everyone is pumped up with Adrenaline. The problem starts happening after you spent significant money and time but you don’t see results. After spending your significant resources, now you want to monetize your product just to sustain itself in the market. But shoot !! You are unable to do it because of the following:</p><h3>Technical Debt</h3><p>I know this reason might not sound convincing. But wait ! Let me prove it.</p><p>Technical debt refers to the accumulated cost of shortcuts taken during the development process of software products. This can include quick fixes, hacks, or temporary solutions that allow for faster development but come at the expense of code quality and maintainability. Over time, technical debt can slow down your product and lead to a variety of negative effects.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*RLkshfqrqrKAT8b8DpJSlQ.png" /></figure><p>One of the main ways that technical debt can slow down your product is by making it harder to add new features or fix existing bugs. As technical debt accumulates, the codebase becomes increasingly complex and difficult to understand. This can make it harder for developers to work on the code, as they may struggle to understand how different pieces of the system fit together. As a result, it can take longer to add new features or fix bugs, which can lead to delays in releasing new versions of the product.</p><p>Developers with less experience are meant to commit these kinds of mistakes. Things get worse if unexpected deadlines are set because in this case, a developer starts using a hack to get the work done.</p><p>The primary solution to this would be,</p><ol><li>Prioritize technical debt reduction: The first step in addressing technical debt is to recognize its importance and prioritize it alongside new feature development. Teams should allocate time and resources to identifying and addressing technical debt in their codebase.</li><li>Conduct code reviews: Regular code reviews can help identify areas of the codebase that are accumulating technical debt. Code reviews can help ensure that code is maintainable, scalable, and well-architected.</li><li>Refactor code: Refactoring involves restructuring code to make it more maintainable and scalable. This can involve simplifying code, removing redundant code, or breaking up large functions into smaller, more modular pieces. Refactoring can help prevent technical debt from accumulating and make it easier to manage existing debt.</li><li>Use automated testing: Automated testing can help ensure that changes to the codebase do not introduce new bugs or errors. Automated testing can also help identify areas of the codebase that may be prone to errors or difficult to maintain.</li><li>Adopt best practices: Using best practices such as code reuse, modular design, and design patterns can help prevent technical debt from accumulating. Teams should also prioritize documentation and knowledge sharing to help ensure that code is maintainable over time.</li><li>Address technical debt incrementally: Addressing technical debt can be a long-term process, so it’s important to tackle it incrementally. Teams should prioritize addressing high-risk areas of the codebase first and allocate time and resources to addressing technical debt on an ongoing basis.</li><li>Get in touch with an Expert: An expert in technical debt has likely encountered similar issues in the past and can bring valuable experience and knowledge to the table. They can help identify areas of the codebase that are prone to accumulating technical debt and provide solutions to address these issues.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/846/1*1L7KkvVA5oKrQ-Offvej_A.jpeg" /><figcaption>We help you with your Startup problem, email us</figcaption></figure><p>An expert can also provide training and guidance to in-house developers to help prevent technical debt from accumulating in the future. This can include best practices for sustainable development, code review processes, and automated testing practices.</p><p>I got an awesome solution for you though!</p><p>We at <a href="https://www.linkedin.com/company/surge-startup/">Surge Startup</a> are a group of experienced entrepreneurs who are willing to help you with your startup journey. We consult startups with their IT technology, product development, and cloud services.</p><p>You can simply DM me on my <a href="https://www.linkedin.com/in/abhishekpatnaik77/">LinkedIn</a> to get in touch if you want your startup’s problem fixed. We at Surge Startup are more than happy to collaborate with your success journey.</p><p>Link to Linkedin:</p><p><a href="https://www.linkedin.com/in/abhishekpatnaik77/">https://www.linkedin.com/in/abhishekpatnaik77/</a></p><p>Or Just Comment “Hi” and we will get back to you.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f933063d8e39" width="1" height="1" alt=""><hr><p><a href="https://medium.com/surge-startup/how-does-tb-cause-startups-to-fail-f933063d8e39">How does TB cause Startups to fail?</a> was originally published in <a href="https://medium.com/surge-startup">Surge Startup</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Building The Failed Facebook Home again in 2021]]></title>
            <link>https://abhishekpatnaik77.medium.com/building-the-failed-facebook-home-again-in-2021-2b50e3a2561d?source=rss-3b21676c7da2------2</link>
            <guid isPermaLink="false">https://medium.com/p/2b50e3a2561d</guid>
            <category><![CDATA[startup-lessons]]></category>
            <category><![CDATA[product]]></category>
            <category><![CDATA[product-design]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Abhishek Patnaik]]></dc:creator>
            <pubDate>Wed, 17 Nov 2021 07:53:32 GMT</pubDate>
            <atom:updated>2021-11-17T07:53:32.803Z</atom:updated>
            <content:encoded><![CDATA[<p><em>What do we learn from the failure of Facebook home in 2013? Well, let&#39;s try planning it again.</em></p><figure><img alt="My approach towards building facebook home in 2021" src="https://cdn-images-1.medium.com/max/1024/1*mDCdK-o4eAP6ye3FKy88lA.jpeg" /><figcaption><a href="https://www.google.com/url?sa=i&amp;url=https%3A%2F%2Fwww.speeli.com%2Fwhy-is-facebook-losing-its-popularity%2F&amp;psig=AOvVaw2mU4vtKe1ehZBZwK7GUD2V&amp;ust=1636987264613000&amp;source=images&amp;cd=vfe&amp;ved=0CAsQjRxqFwoTCKDAgPuKmPQCFQAAAAAdAAAAABA0">Image Source</a></figcaption></figure><blockquote>On April 4, 2013, Facebook invited media to “Come to See Our New Home On Android”. Sources tell us it will be a modified version of the Android operating system with deep native Facebook functionality on the home screen that may live on an HTC handset. The evidence aligns to say this is the Facebook Phone announcement people have been speculating about for years.- <strong>TechCrunch</strong></blockquote><p>Facebook took a big decision of launching Facebook home back in 2013. For those who don&#39;t know about the Facebook home, it was supposed to be a default home screen for mobile devices. The major business goal was to increase user engagement.</p><p>Having a home screen on android would help users use the Facebook app more frequently, thereby increasing user engagement.</p><p>It was available on some high-end HTC and Samsung phones, and the HTC First phone to be completely optimized for Facebook Home.</p><ol><li>What was the major problem with Facebook Home?</li><li>Agitation</li><li>Outrage | Media Attention</li><li>Solution in 2021</li><li>Risk Assessment</li></ol><h3>What was the major problem of Facebook Home?</h3><p>The home shows facebook’s social updates, contacts, messaging services, most importantly, Facebook ads directly on an Android user’s home and lock screens. The home was supposed to make Facebook more easily accessible even without opening the app. Isn&#39;t it awesome to be able to see every single update from your friends without doing anything?</p><p>Of course, it would be. But that&#39;s not all everyone needs. Taking the customer&#39;s segments into consideration, not everyone is a Facebook addict. I mean, personally, I didn&#39;t even use Facebook much until I had my ads running for my page. So there raised a problem of not being able to see the desired app. This problem can be solved efficiently in 2021 though(will discuss it in the end).</p><p>Trying to change the existing way users used Android was another big issue. Users were not really ready to make a change in the conventional workflow.</p><h4><strong>Entering into a saturated market</strong></h4><p>I would never like to travel on a train that has no space or that&#39;s too crowded until it delivers some extraordinary results. The same goes with products, there were already many launchers on the play store that was good enough. So this was not something of a major pain point for a user.</p><h4>Agitation:</h4><p>The Facebook home was an awesome feature for those who used it often, but it became a really tough time for those who wanted to use other apps. People found it really difficult to find another app, which created a disconnect between the vision vs the use case.</p><p>That was not it, people now wanted to uninstall it as the user experience was not that efficient anymore. An article from <a href="https://www.theverge.com/2013/4/12/4216926/how-to-get-facebook-home-or-uninstall#:~:text=GETTING%20RID%20OF%20FACEBOOK%20HOME">theverge.com</a> gave the internet a clear view of the non-effectiveness of the update.</p><figure><img alt="agitation due to facebook home" src="https://cdn-images-1.medium.com/max/817/1*f2PeggPWlsrRopBsDX86Pg.png" /></figure><h4><strong>A world outside Facebook</strong></h4><p>Facebook is an awesome tool. It&#39;s of many use. But when the home came, they tried to capture the market they didn&#39;t aim to. It was too much Facebook for people who didn&#39;t use it much. The problem started arising because it became too “much sugar in a coffee”. This goes well with the saying “Excess of everything is harmful”. With the new home, all users saw was facebook and apps around it. It was too much for users to digest.</p><h3><strong>Proposing a solution for Facebook Home in 2021</strong></h3><p>We talked about a lot of problems and agitation. Now let&#39;s discuss the solution. If I had to launch Facebook home again in 2021 this would be my approach.</p><p>The first step would be to decide the aim of the launch. Since the initial goal was to increase user engagement I would keep it that way.</p><p>So keeping user engagement in mind, the next step would be to decide the customer base. I would like to divide the customer base into two parts, the first one for personal use and the other one for enterprise use.</p><p>The ideal customer base would be those who use Facebook for non-business use. Whereas for business/enterprise usage, I would consider the customers who use Facebook for running ads and would like to stay updated with every bit of information.</p><p>So let&#39;s start with the personal usage Facebook home screen. Taking customer empathy into consideration I would set goals as “Making customers achieve results faster”.</p><p>I would define my customer journey in the following steps:</p><ol><li>The user opens his/her mobile phone and sees the homepage</li><li>The user finds the app using slides or searches from the home screen</li><li>Select the apps, starts using it</li><li>After using it for a few mins, the user might switch to some other app.</li><li>He/she might come back to the home screen to search for another app, and the loop continues.</li></ol><p><strong>The important thing to notice here is how easily the user finds the desired app.</strong></p><p>I would consider the following steps to increase the usability of The Home:</p><ol><li>Use data to prioritize different apps on the home screen, this would thereby reduce the number of clicks to reach the desired app</li><li>Allow users to star/pin different apps on the home screen</li><li>Provide a Top search bar to search and thereby sort considering the analysis done at step 1</li><li>Sort the Apps screen with the most used apps at the top</li><li>Show other apps like Instagram and WhatsApp</li></ol><p>The next step would be to prioritize the features. Not all of the above would be useful but let&#39;s try to prioritize the best assumption we can do.</p><p>Let&#39;s create a comparison table and try to understand it better. I would start by assigning a point system to each feature. The best idea has a score of 5.</p><figure><img alt="score card for different metrices" src="https://cdn-images-1.medium.com/max/668/1*J16vJyUpSpxwQ-pJlxrHnQ.png" /></figure><p>Explanation of the scores:</p><ol><li><strong>Data Analysis to Prioritize app</strong>: Facebook has a ton of data. They can use it efficiently to find the behavior of the users and build an AI around it to predict the right time considering the time, demographics, and semantics of the user. If this works well, it would be a complete game-changer like the autocomplete google keyboard. Let&#39;s say I am standing near a street trying to find a cab. Being able to do the right analysis and showing recommendations for cab apps would be really cool! But this comes with a risk factor, which would include a lot of time to build and test. So that is the reason for the high risk and effort score.</li><li><strong>Allow Pin/Star Apps:</strong> Pinning apps on android is a default feature in android. So I assume it won&#39;t require effort like the previous one. So give it a medium to a low score of -2.</li><li><strong>Providing a search bar: </strong>Being able to index the app&#39;s information and finding out the right app using a search functionality would be easy. As it requires soft matching of text. That is the only reason for low scores for this feature.</li><li><strong>Sort the App on Recent Used app: </strong>Being able to see the most recently used apps would save time for users. Gave it a comparatively high score for efforts as this might need comparably more effort that.</li></ol><p>Anyways the efforts section may change with consultation with the engineering team.</p><p>I would prefer to go with ideas 2,3 and 4. <strong>Idea 1 is a game-changer but might not fit well in the MVP version for the product</strong>. So I might keep it for the future roadmap and start implementing it after we have enough data.</p><figure><img alt="proposed solution and UI" src="https://cdn-images-1.medium.com/max/519/1*N8yklpoPI88F2JMv4FiSsA.png" /></figure><p>We spent a lot of time discussing the pain points and solutions. Which was important to solve because that was the major flaw in the previous version.</p><p>Now let&#39;s talk about Facebook apps. Let&#39;s keep it simple and introduce a facebook section on the home screen. This facebook section shows the most used Facebook apps. We also have a small updates section that will serve the purpose of showcasing new updates from Facebook. Thereby increasing the engagement.</p><p>We preferred having only a single section for facebook because this was the main pain point of users. So that needed to be addressed properly. Apart from that, we could also use the notification system more efficiently.</p><p><strong>Building for Enterprise Customers</strong></p><p>It becomes quite necessary for enterprise customers to stay updated with there the ads they are running or any important notification regarding their business account. Often it happens that ads stop unexpectedly due to some issues, so this would be the right opportunity to show those messages. This would keep the enterprise with smooth conversion as well as this will also add on to the revenue.</p><p>There&#39;s a lot more we can do with enterprise or business customers on Facebook. Comment down if you want to see the 2nd part of this blog where we discuss more the business customers.</p><p><strong>Risk Assessment:</strong></p><p>There’s definitely a risk involved in bringing back facebook home again. That being the sole purpose, we choose options 2,3, and 4 as the risk involved was less. This could come as an inbuilt feature inside the Facebook app itself in the new versions. Users could easily locate the update when they first time opened the app(after the update) or in the home settings for the phone.</p><p><strong>Conclusion:</strong></p><p>Often simple things work the best. There has been numerous example where just a single thing worked the best. We kept customer empathy as the top priority. While working on the user engagement. Launching an MVP to beta testers and leading it via the learn-build-measure loop would result in a promising update.</p><p>If u did like the article then consider following me up on Medium.</p><p>Let&#39;s connect on LinkedIn to see if we have similar interests.</p><p><a href="https://www.linkedin.com/in/abhishekpatnaik77/">Abhishek Patnaik - Product Manager - Quantamix Solutions | LinkedIn</a></p><p><strong>References:</strong></p><p><strong>The Verge</strong>: <a href="https://www.theverge.com/2013/4/12/4216926/how-to-get-facebook-home-or-uninstall">https://www.theverge.com/2013/4/12/4216926/how-to-get-facebook-home-or-uninstall</a></p><p><strong>Quora</strong>: <a href="https://www.quora.com/Why-was-Facebook-Home-received-poorly">https://www.quora.com/Why-was-Facebook-Home-received-poorly</a></p><p><strong>The Business Insider</strong>: <a href="https://www.businessinsider.in/tech/facebook-home-is-a-flop-employees-know-it-and-users-dont-like-it/articleshow/21120024.cms">https://www.businessinsider.in/tech/facebook-home-is-a-flop-employees-know-it-and-users-dont-like-it/articleshow/21120024.cms</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2b50e3a2561d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Most PM’s/Founders underestimate the power of determining a Niche]]></title>
            <link>https://abhishekpatnaik77.medium.com/most-pms-founders-underestimate-the-power-of-determining-a-niche-7af2b446b2c2?source=rss-3b21676c7da2------2</link>
            <guid isPermaLink="false">https://medium.com/p/7af2b446b2c2</guid>
            <category><![CDATA[enterpreneurship]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[mvp]]></category>
            <dc:creator><![CDATA[Abhishek Patnaik]]></dc:creator>
            <pubDate>Mon, 11 Oct 2021 04:30:15 GMT</pubDate>
            <atom:updated>2021-10-11T04:31:03.770Z</atom:updated>
            <content:encoded><![CDATA[<p>Ever built a product without having a clear vision of whom to sell? <strong>Congrats you are not alone!</strong></p><p>In this post, we will talk about how is niche discovery an important part of a business.<br>This is one of the most common patterns that founders or PMs follow in order to make their products.</p><h4><strong>Why is it important to get started with a niche?</strong></h4><p>Let&#39;s start by understanding why is it necessary to Determine your niche.<br>Before that, you need to do one thing. Close your eyes and tell me what do you see?</p><p>Nothing??? that is exactly how building a product is without defining a niche.</p><figure><img alt="niche selection" src="https://cdn-images-1.medium.com/max/780/1*T3bZNTeIrGYu3--4WXqrJQ.png" /><figcaption>Different paths to get started from (<a href="https://www.designed-destinations.com/wp-content/uploads/2018/07/dd-intersection-blog.png">source</a>)</figcaption></figure><p>While building a product it is always a good idea to start with a small group of people who will be your ideal buyer persona. Example: <em>Apple is building an iPhone for kids</em>. Apple might start by narrowing down the audience in order to find the right audience and start by building the phone.</p><h4><strong>Why is it a good idea to start with a small group of people?</strong></h4><figure><img alt="limited sources in startups" src="https://cdn-images-1.medium.com/max/525/1*DsbD8FSnFRLfAPlbabiOJA.jpeg" /></figure><p>A small group of people is generally are called ideal customers. It is important to focus on a single buyer persona because of the following reasons:</p><ol><li>Startups are limited with resources and time, so can&#39;t focus on 100s of things at once</li><li>Focusing on a single buyer persona helps you create things faster, thereby helping to validate the right product-market fit.</li><li>You know the right path and improvise things depending upon the feedbacks.</li></ol><h4>Do we lose the bigger audience while focusing on a single buyer persona?</h4><figure><img alt="starting with a small niche" src="https://cdn-images-1.medium.com/max/900/1*wqhSsZ2Dsruaq_HqZOIDTA.png" /><figcaption>Start with a single niche, then focus on a bigger audience</figcaption></figure><p>No, we don&#39;t. The concept goes as we start with a small group of people and then keep expanding after we satisfy the initial customers.</p><p>You gain trust when you satisfy the feedbacks of the initial customers. The initial customer then server as your reviewers. They recommend the tool to others, and this is how the tool grows.</p><p>So this is why it&#39;s necessary to start with one buyer persona and then grow with their trust and feedback.</p><p>Let&#39;s take an example and try to understand:</p><p>Let&#39;s try creating a new feature for the iPhone for creating fancy filters like snap chat.</p><p>Now there&#39;s a wide variety to choose from, so which one to go with??</p><p>From the problem statement, it is clear that we are not focusing on android users. Now we can choose from a wide variety of persons like children, teenagers, couples, travelers, and many more.<br>Filter’s are most likely to be used by teens, who like clicking selfies with friends and family. So let&#39;s define our ideal buyer persona as</p><p>“Tina a 17 years old teenager who loves clicking selfies with her friends”</p><p>Well, now we have a clear idea that we will be building an app for a teenage girl. So let&#39;s say that the Niche is my friend. I know everything about her. And I know that she likes to watch Anime, she loves Hollywood star Tom Cruise, she loves swimming.</p><p>This now gives me a clear idea of what I need to build. We talked about building an app for filters, so what are those filters that Tina might love.</p><p>maybe a filter that converts her face to anime characters?</p><p>maybe a filter that lets her click selfie with tom cruise?</p><p>maybe a filter for swimming near a shark?</p><p>Possibilities are endless but when we are talking about a single niche, we start to narrow down the possibility, thereby creating a tool that users love.</p><p>As from the above example, we just saw, when we considered a single entity how much easy it became for us to give the product a vision. Giving a vision to a product is important for a product manager or founder.</p><h4>“Can my brother use this?”</h4><p>These are some of the most common questions we see on product review websites. But what does it exactly mean?<br>Well, they don&#39;t mean their own brother, what they mean is a young middle-aged boy, he can be a pilot, he can be a teacher, what it means exactly is can this app be used by a mid-aged man, who is not that good that tech-savvy.</p><p>Looking into basics Product managers need to be good spies. They need to know the best pain points of their buyer persona.</p><p>This is the most important part to address, who is the company solving the problem for.</p><p>Good personas will have a fictional name, in our case we named it Tina.</p><h4>Overwhelming Buyer Personas</h4><figure><img alt="one niche at a time" src="https://cdn-images-1.medium.com/max/900/1*3Ur2rkm0XZz4uRmY2WryAQ.png" /><figcaption>Keeping things simple is important</figcaption></figure><p>It is very tempting to make a wiki of your buyer persona. But at the same time, it becomes overwhelming. Keep the things detailed but don&#39;t overflood with data. It would be a good idea do to define your persona with a detailed statement. Yet if you think it&#39;s not relevant to your product don&#39;t include it.</p><blockquote><em>Eg: Tina is a 17 years old teenager who loves filters on her iPhone. She loves to dance, ride horses.</em></blockquote><p>In this case, she loves to dance and ride horses are not related to creating filters. So I won&#39;t consider this to into my buyer persona. I want to have crisp information that helps me focus on my niches desires, pain points, and solutions. For example, a developer is unable to enjoy his weekend due to the deployments, so why not create an automated deployment pipeline that helps him rest on weekends.</p><h4>Building Your Own Custom Niche</h4><p>The right way to visualize your ideal customer would be to be super-specific. That being said, let&#39;s not forget the psychographics and emotional side. These two act as a catalyst to increase your conversions.</p><p>The best example for this would be Supriya who wants to lose fat so that she can become Mrs.Asia. Now that we know her priority, we can build our tool accordingly. Our use case is a combination of a pain point and emotional side.</p><p>Demographics are important, but not in all cases. For example, Flipkart trying to launch a new country would need to target the temperature, climate, economy, and their usual clothing patterns. A person staying in the colder side of the country would need warm clothing, he would seldom look for shorts or pajamas. On the other hand, a person staying near the beach would prefer some shorts or pajamas because of the humidity and breeze.</p><p>It is not always possible to have prior information about customers. We can start off by creating a rough persona, then validating and refining it throughout the product life cycle.</p><h4>People Don&#39;t buy the product they buy the solution</h4><p>Everyone likes to make their work easy. We look for solutions that solve a major problem. To make sure that the solution is on the right track it is important to keep validating the product.</p><figure><img alt="customer success" src="https://cdn-images-1.medium.com/max/830/1*-NL7J6-Ysl87CcRSWtJ7KQ.png" /></figure><p>A product should be able to add value to its customer&#39;s journey. It is easy to build something complex, but it is very hard to build something that is simple yet effective.</p><p>The only way to show love towards a customer is by solving feedbacks from your buyer persona. We hire content writers for their ability to write content. A Content writer helps his/her clients to write an article based on some seed inputs. They then build the complete article for you depending on those inputs. Everyone can write content, then why do we need to hire content writers? Well, there are many factors like time, skills, and of course efficiency. Content creates provides a wide range of specialties in researching and writing content for you. This saves time, the client gets a copy that is highly researched, SEO optimized, and reliable.</p><p>Everyone wants a product that helps them live a better life, everyone wants a better version of themselves in a tool.</p><p>Validating the idea for future roadmaps with customers stays an important part of the product lifecycle. We may sometimes end up creating a wrong buyer persona, for that it&#39;s important to keep validating the idea.</p><p>This can be also said that people buy a different version of themselves.</p><p>Here&#39;s an awesome article put to by <a href="https://medium.com/startup-grind/people-dont-buy-products-they-buy-better-versions-of-themselves-2ce85fdb5ff1">Startup Grind</a>.</p><p>Creating a persona and letting it drive your business is a proven method to lead the business.</p><p>I have been into the product for about 4 years. Always loved to learn new things. If you made it to the end, let&#39;s connect on LinkedIn. Would love to hear your views/feedback.</p><p><a href="https://www.linkedin.com/in/abhishekpatnaik77/">Abhishek Patnaik - Product Manager - Quantamix Solutions | LinkedIn</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7af2b446b2c2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[5 ways to screw your sleep as a Developer]]></title>
            <link>https://abhishekpatnaik77.medium.com/5-ways-to-screw-your-sleep-as-a-developer-1476e165b150?source=rss-3b21676c7da2------2</link>
            <guid isPermaLink="false">https://medium.com/p/1476e165b150</guid>
            <category><![CDATA[development]]></category>
            <category><![CDATA[healthy-lifestyle]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[health]]></category>
            <category><![CDATA[developer]]></category>
            <dc:creator><![CDATA[Abhishek Patnaik]]></dc:creator>
            <pubDate>Fri, 30 Jul 2021 07:26:19 GMT</pubDate>
            <atom:updated>2021-07-30T16:33:21.022Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>“It’s a badge of honor to be a workaholic. But I don’t want to die at my desk”</blockquote><blockquote>- A developer</blockquote><p>Developers are very passionate about their work, so they often find themselves working on projects late into the night.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ffj25Nib2n3SIg7THZ743A.jpeg" /><figcaption><a href="https://casper.com/blog/sleep-anxiety-guide/">Image source</a></figcaption></figure><p>To be a programmer is a noble profession. To be a good programmer is even nobler. I’m not talking about how many lines of code you can write in an hour, or how many bugs you can fix in a day, or how much time you spend on Stack Overflow. I’m talking about the impact your work has on the world around you and your ability to solve complex problems with simple solutions.</p><p>As a developer and a product owner, I have realized that it is very important to have a proper sleep while you code.</p><h4><strong>Possible reasons for a developer to not get proper sleep.</strong></h4><p><strong>a. Doesn&#39;t follow good standards: </strong>Problems are meant to arise in development if good standards are not followed.</p><blockquote><em>But what does good standards have to do with sleep?</em></blockquote><p>Well, a lot. When you write shit code it reflects in your personal life. I will give you an example from my personal life. Back in 2017 when I started my career as a data scientist. I was moved away by the tremendous speed of every developer. Since it was a startup, everything was expected to be delivered ASAP. During the initial month, I was busy learning, I wasn&#39;t able to serve any deadlines.</p><p>Due to less experience in development, I was not able to write production-level code. I followed my own standard called “Make the code working”. It doesn&#39;t really make sense.</p><p>If you don&#39;t have good standards, code is meant to fail several times. We then sit to debug the code and make it work “anyhow”. <br>With my experience in development, I have seen developers write “hacky” code which thereby leads to more fixes, which there leads to more hours of codes, which there leads to <strong>more stress and less sleep</strong>.</p><p>When devs follow no standards it takes more time to debug the code. <br>Here is a great article on how important standards are in programming: <a href="https://www.multidots.com/importance-of-code-quality-and-coding-standard-in-software-development/">link</a>.</p><p><strong>b. Using 3rd party API blindly: </strong>I wish development was that simple. We see memes of copy-paste from open source but little do we understand the use case. We simply copy-paste or directly use the 3rd party API.</p><blockquote>Note: This doesnt refer to API’s from well reputed organizations.</blockquote><p>It is not really a good practice using a 3rd party API “workarounds” in your production. In case you have it, start thinking of an alternative. We never know when the 3rd party stops working or their company shuts down.</p><p>Now, this doesn&#39;t mean that if we should not use 3rd party APIs. It is fine in the nascent stage of a startup, but as the startup grows it&#39;s a good idea to think of an alternative.</p><p>Example: In one of my startups we used a 3rd party API to crawl the websites. One morning that website came down, we later came to know that the startup was shut down. Users went crazy as most of the things were not working. We had to immediately find some other way to fix that issue.</p><p><strong>c. Not understanding the business value: </strong>Generally it&#39;s the team lead/managers to make sure that the expectation is set. Expectations mean setting clear goals and designs.</p><p>Without understanding the business value it&#39;s really very hard to build anything. Understanding the business aspects of development helps developers build things faster and with a clear vision.</p><p>Apart from that more importantly it helps developers understand the use case and build the code accordingly. Taking various use cases into consideration helps users build code accordingly thereby handling more exceptions.</p><p>If use cases are not considered then it becomes hard to handle the exception thereby more errors leading to more stress and less sleep.</p><p><strong>d. No time management: </strong>Problems start when we don&#39;t have control over our time. Time is something which is very costly so it must be spent wisely. Developers generally lack this habit. Especially this WFH has made the work of time management more difficult.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/700/1*puN5kFkYspZN-xoBGGBELg.jpeg" /><figcaption><a href="https://www.google.com/url?sa=i&amp;url=https%3A%2F%2Fwww.bamboohr.com%2Fblog%2Feffective-time-management-activities%2F&amp;psig=AOvVaw1tKvwWYPuhjKlPp4s5-h5t&amp;ust=1627620725056000&amp;source=images&amp;cd=vfe&amp;ved=0CAsQjRxqFwoTCOiSk6bFh_ICFQAAAAAdAAAAABAL">Image Source</a></figcaption></figure><p>There should define boundaries that separate work and personal life. Without that, it makes life a pretty mess.</p><p><strong>e. Working on 10 things together: </strong>Devs tends to work on 10 different things together without completing one properly. This thereby results in too much pressure. Patience is the key, yet most of us lack this. Though is the duty of both developers and team leaders to make sure that devs don&#39;t do multiple tasks at once.</p><h4>Problems Due to less sleep:</h4><p>A study from the American Journal of Preventive Medicine found that more than 10% of the world’s population is suffering from insufficient sleep. This lack of sleep has been linked to a variety of health issues, including diabetes, obesity, and depression. In addition to these common consequences, there are also specific effects on developers: Developers who suffer from sleep deprivation have lower job performance and higher rates of mental errors in their work. They also produce fewer lines of code in a given time period than those who get adequate restful sleep.</p><p>Poor sleep can reduce your life expectancy. It has been shown that a person’s health is inversely proportional to the amount of sleep they get.</p><p>Apart from that depression and anxiety are two mental health disorders that can be caused by lack of sleep. This is because a lack of sleep can lead to the release of certain hormones that cause people to feel more anxious or depressed.</p><p>Here is a link to how important sleep is: <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4849507/#:~:text=Many%20studies%20have%20shown%20that,diabetes%2C%20stroke%2C%20and%20obesity.">Click here</a></p><h4><strong>Making sure that developers get proper sleep</strong></h4><p>Avoid screens an hour before bed and have a routine that lets you be ready to sleep at night.</p><p>Explore new things, life is much more than a cubicle. Travel, meet new people, spend time with family.</p><p>Hustle when you need to, but make sure that you hustle in the right way.</p><p>One more thing if the above doesn&#39;t help to get sleep then consult a mental health practitioner..hehe.</p><p>Well, that was all from the side. If you have made it to the end, don&#39;t be shy to hit the like button.</p><p>Comment your views. Would love to know your views.</p><p>Let&#39;s get in touch on LinkedIn:</p><p><a href="https://www.linkedin.com/in/abhishekpatnaik77/">Abhishek Patnaik - Product Manager - Quantamix Solutions | LinkedIn</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1476e165b150" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[5 Reasons Why Your MVP Might Not Work]]></title>
            <link>https://abhishekpatnaik77.medium.com/5-reasons-why-your-mvp-might-not-work-182736941e83?source=rss-3b21676c7da2------2</link>
            <guid isPermaLink="false">https://medium.com/p/182736941e83</guid>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[mvp]]></category>
            <category><![CDATA[productivity]]></category>
            <dc:creator><![CDATA[Abhishek Patnaik]]></dc:creator>
            <pubDate>Sun, 18 Jul 2021 10:55:11 GMT</pubDate>
            <atom:updated>2021-07-18T10:55:11.963Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3FTMA-lAy_mLL6nvazJ7Ng.jpeg" /></figure><p>90% of startups fail due to a lack of planning on an MVP version. Certainly, all ideas are worth coming to market but they don&#39;t really hit the benchmark because of a bad MVP version.</p><p>Creating an MVP depends on many factors like</p><ol><li><strong>time duration to complete the MVP:</strong> The deadlines or time duration to set MVP affects a project deeply. I have seen project managers, CEO fall prey to this loophole. They tend to set unachievable deadlines for the team which results in a lack of research, competitor analysis, and bad planning.</li><li><strong>Resources:</strong> A startup always has a problem with resources, so it&#39;s very necessary to plan your MVP depending upon the resources. By resources we mean: developers, computational limitations, financial earnings.</li><li><strong>User Basic requirements:</strong> There are some basic sets of features we decide to provide to users initially. We cant go with all features so we decide on a limited set of features to provide to users. These features are decided based upon the target niche, their demographics, psychographic, competitors, and many more.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3RQHNylcIx1AOeYgfl6SwQ.png" /></figure><p>Here are Top 5 reason why your might fail.</p><ol><li><strong>Lack of planning: </strong>Planning is important in nearly everything we do. Even a Family Trip cant be successful without a proper plan. You see, how important planning is right !! This becomes more and more crucial when it comes to building an MVP version. <br>There are lots of things that need to take into consideration while planning.</li><li><strong>Pulling too many eggs into a plate: </strong>Often pulling too many features into a single go leads to brutal fallback. There&#39;s a saying you can either build 100 features and close them all at once or build a few and evolve them.</li><li><strong>Trying to leap a mountain with a single jump:</strong> Are you the one who wants to achieve overnight success? Congrats, you are not alone. Startups need a lot of patience. No one gains overnight success. You would have heard stories of apps gaining success overnight, but less do we know about the failures the founders would have gone through.</li><li><strong>Oops!! You missed users feedback:</strong> User feedbacks are very important as they help in shaping the right future of an app. User feedback stays at the top priority in any startup. With time we have experienced that founders, entrepreneurs with great goals tend to neglect user feedback at some places.</li><li><strong>Unprofessional team: </strong>MVP comes with high risk and responsibility. Team really needs to be focused with work. It might include staying awake for hours. Following are signs of a unprofessional team:</li></ol><p><em>a. Taking deadlines lightly</em></p><p><em>b. Executing stuffs without a proper plan or discussion</em></p><p><em>c. Poor quality of work</em></p><p><em>d. Having a high EGO</em></p><p><em>e. Not learning from there mistakes</em></p><p>Well <strong>being unprofessional also conveys the same for managers</strong>. Managers sometimes fail to bring out the right message to there team members. Being toxic doesn’t mean you are a good manager. Taking right decision at right time is what managers are meant to do.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DiCv9-KyarSrG13LtqGkDw.jpeg" /></figure><h3>I can do 10 things and end no where, or try one thing and take it to places.</h3><p>Its not always a good idea to put legs onto two boats together. In real life we sink if try to do so. Same happens in startups we tend to drawn when we have too many things, but nothing makes sense.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LLydXaU0dhKExqdf4aSA1A.jpeg" /></figure><h3><strong>Knowing your Customers more than there Family does</strong></h3><p>It is very necessary to know your niche before you start building anything. I have always loved the concept of understanding my ideal buyer persona.</p><p>Just an short example for the same, We at CrawlQ narrowed down the tool, just to make sure that it was relevant for our niche. Initially CrawlQ had the following features:</p><ol><li><em>Market Research</em></li><li><em>Competitor Analysis</em></li><li><em>Topic Modeling</em></li><li><em>Content Automation</em></li><li><em>SEO</em></li></ol><p>We at CrawlQ promise to be a content automation tool. Being a startup we were not able to focus on soo many features. So we thought of “<strong>UNHAND</strong>” few features.</p><p>Given limitations in resources and time, it was very difficult for us to focus on so many things. We thought of sticking only with Market research and content automation. We started adding more value to it by adding new updates, improving UX UI, making it more and more usable for content automation.</p><p>Users loved our tool more after that.</p><p>Now the question is, <strong>Why did we do so?</strong></p><p>We focused on our “Niche”. We knew our ideal buyer persona were B2b SaaS founders, Entrepreneur, Solopreneurs.</p><p>This comes with a risk of wrong interpretation, But there’s no fun without risk.</p><p>Risk is fated part of life, be it a business or normal life. Risk will always be there. But to build a MVP product with right decision and aspects will always boast your sales by 10 times. This method also helps in lesser failure rate.</p><p>If you loved reading the article, please consider clapping and sharing it.</p><p>Lets connect to discuss more on trending topics.</p><p><a href="https://www.linkedin.com/in/abhishekpatnaik77">https://www.linkedin.com/in/abhishekpatnaik77</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=182736941e83" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How did our startup earn $70K revenue in just 2 months?]]></title>
            <link>https://abhishekpatnaik77.medium.com/how-did-our-startup-earn-50k-revenue-in-just-2-months-d0eed81f9045?source=rss-3b21676c7da2------2</link>
            <guid isPermaLink="false">https://medium.com/p/d0eed81f9045</guid>
            <category><![CDATA[growth]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Abhishek Patnaik]]></dc:creator>
            <pubDate>Wed, 19 May 2021 05:35:09 GMT</pubDate>
            <atom:updated>2021-11-17T11:14:31.459Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0aCK7wXcMZeled59_oL4dQ.png" /></figure><p>Everyone’s talking about raising money via sales, right? We all know that “raising money”. But when you learn it and use it, you’ll see a huge difference in your raising money via sales and how fast you can get more money in a startup.</p><p>In this blog, I will share my experience in my startup that has led us to earn more than $70k in just two months. These are the possible questions we will discuss: <br>What does the content cover?<br><br>how does your target audience want to see you?</p><p>Why do startups need patience?</p><p>Let&#39;s start with the basics of who we are and what we do?</p><p><strong>CrawlQ’s</strong> AI-powered Content Automation Tool helps Founders, Marketers and Copywriters to Profile Ideal Customers, Differentiate from Competitors and Optimize Content for Sales, SEO &amp; Social Media.</p><p>We were never a Content automation tool, we started off initially with SEO and topic modelling for websites.</p><p>We did go through a <a href="https://ardas-it.com/full-guide-to-successfully-pivot-your-saas-startup#:~:text=What%20Is%20A%20Pivot%20in,business%20model%2C%20or%20growth%20engine.">Pivot</a> which made us an SEO optimized content automation and Market Research tool.</p><p><strong>So what did we do exactly?</strong></p><h3>We took our time understanding the Market</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/620/1*djNhD2qOnvBp9w0gOT8R_w.jpeg" /></figure><p>We did take our time understanding the market. The product is about 2 years old now. But it was in last August we decided to hit the market with whatever we had. We started building everything around content automation. We had a Sales Copywriting built for us, but we did more to be called a Content automation tool. So we came up with an idea for an AI content writer.</p><p>That was just a start, we now have almost all of your steps automated.</p><p>So by December, we were all set to launch. But oops !!!</p><h3>Oops !! we forgot about user experience</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/300/1*Tsda2r9pC7tQMLT6C4EFGQ.jpeg" /></figure><p>Yeap I know, many of you will say “Come on Abhishek, UI UX is the last thing for a product”.</p><p>Yeah yeah, I know. But we had this strong feeling that if we continued with the existing flask app it would be pretty difficult to give users a better user experience in future.</p><h3>Transition Stage</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/577/1*C3IIoy2Cb1S70JfS70T2ZA.jpeg" /></figure><p>Now with everything set in Flask, we had to move to react. We just wanted to make sure that we have the upper hand on making changes for UX. And we can do it whenever we want.</p><p>I still remember the days, we had to create more than 90 endpoints. Though our team achieved it within a month. Our react app was ready by 15th Feb. We had to run fast though. As fast as we could……….</p><h3>Please wait, the App is not ready yet</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/404/1*azEYvzCfiLDmLvKKytYCzw.png" /></figure><p>I still had a strong feeling that the UX was not good. We needed more time to improve it. While I was trying to make it better and push the deadline. Our Founder <a href="https://www.linkedin.com/in/harish-kumar-frm-75427116/">Harish Kumar</a>, said if we keep doing this believe me we will never be ready. So let&#39;s hit the market and try to analyze what happens.</p><p>Harish had a valid point, and so we settled on not moving the deadlines further.</p><h3>Not Ready….Yet not afraid to fail.</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/387/1*mwpSAUO6Fi5xheQntqiwAw.jpeg" /></figure><p>Well, all we wanted was to hit the market and see how things work in reality. So finally we did launch on 8th March on <a href="https://appsumo.com/crawlq-ai-content-automation/">AppSumo</a>. Our heartbeat went crazy that day. We knew what was coming.</p><h3>Build-Measure-Learn</h3><p>When we hit the market we got an overwhelming response from users and We are really glad to have it. We started getting feedback from our users. And that is what we exactly wanted. We took a deep breath and started working on our user&#39;s feedback. We kept users feedback at top priority.</p><p>We built new features, updated them, measured our progress with users feedback and learnt from it.</p><h3>Building our CrawlQ Community</h3><p>My friend Harish Kumar always had this awesome idea of creating a community around our tool. We did take our time creating our own community. Just after the launch, we had our own <a href="https://www.facebook.com/groups/crawlq">Facebook group</a>.</p><p>We got tons of emails, few were for feedbacks, few had some issues, few wanted to add more feature.</p><p>Gladly, we had a marketing team that was really patient in replying to each and every issue. The marketing team was really patient in solving issues, talking to the team and adding improvement/new features into the backlog.</p><h3>Patience</h3><p>This is the most underrated yet the most effective key that helps any startup grow. If you lose patience and say, “OK I m done with this. Maybe I was never made for a startup”. You would never realize you were just so close to making a good amount.</p><h3>Customers are Everything</h3><p>We focused all our energy on trying to understand what users want. We still are doing it. We will keep doing it. And we are in love with this process.</p><p>Our initial progress was slow. But we took our users feedback, worked on it. As we worked on users feedbacks, we started creating a difference.</p><p>Users themselves came up with ideas that is actually helping us to improve the product more and more. We were lucky enough that some users voluntarily came up, shared there ideas.</p><h3>What actually did the market lacked?</h3><p>Users wanted a tool that could generate long-form content based upon a deep understanding of their niches.</p><p>But to generate that kind of content we needed data from users. We already had a market research tool, that helps the user in understanding the market. So why not use the same data?</p><p>We started using the same data and started generating content based on those niches. Users started loving insights about their Niches, this gave them an idea about the market. Helping them understand the market deeper.</p><p>According to users request, we also added new features for generating headlines, meta-description and blog outlines for our users.</p><p>We still are adding more features to our application making sure users can get more of content automation.</p><p>We would love to get your views on it.</p><p>I would personally like to thank a close friend <a href="https://www.linkedin.com/in/mahaveer-muttha/">Mahaveer Muttha</a>. I followed his feedback approach and findings of the market and it helped a lot.</p><p>If you made it till the end, then why shy pressing the <strong>like button, commenting and sharing</strong>…haha. Anyways thanks for reading this article so patiently.</p><p>Now, here is a link to our app. Are you the one who wants to buy a tool that makes content based on your niches or want to do awesome AI-based market research?</p><p>Or are you simply someone who is looking for AI content automation?</p><h4><strong>CrawlQ is the one place stop for you</strong>: <a href="https://quantamixsolutions.com/pricing/">Visit our page</a>.</h4><p>We have a team who is always ready to hear your feedback, mail us at support <a href="/service/mailto:support@quantamixsolutions.com">support@quantamixsolutions.com</a> if you have any queries.</p><p>Want to convert your website into an organic growth machine?? <a href="https://quantamixsolutions.com/how-to-turn-your-website-into-an-organic-growth-machine/">Click here</a></p><p>Get in touch with us, below is our social handles.</p><ul><li><a href="https://www.linkedin.com/in/abhishekpatnaik77/">Abhishek Patnaik - Product Manager - Quantamix Solutions | LinkedIn</a></li><li><a href="https://www.linkedin.com/in/harish-kumar-frm-75427116/">Harish Kumar, FRM - Prodct Owner - Rabobank | LinkedIn</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d0eed81f9045" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Django Vs Flask. The End.]]></title>
            <link>https://abhishekpatnaik77.medium.com/django-vs-flask-the-end-fb9c33ceb4ea?source=rss-3b21676c7da2------2</link>
            <guid isPermaLink="false">https://medium.com/p/fb9c33ceb4ea</guid>
            <category><![CDATA[python-programming]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[django]]></category>
            <category><![CDATA[flask]]></category>
            <category><![CDATA[software-development]]></category>
            <dc:creator><![CDATA[Abhishek Patnaik]]></dc:creator>
            <pubDate>Sun, 20 Sep 2020 13:54:00 GMT</pubDate>
            <atom:updated>2020-09-20T13:54:00.208Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rAjo2V2bQ2nJIQu7hv6KUg.png" /></figure><p>This has been the most controversial topic among flask and Django developers. There has been a war waging since both came into existence. So, let&#39;s find out what’s the difference between both of these packages.</p><p>Let&#39;s start from basics,</p><h3>Why is there a controversy between the two?</h3><p>Both frameworks do the same work of serving for building a web-application. For example, there&#39;s always a controversy between i-phones and Andriod phones. Both smartphones do the same work, but yet the ways of doing it are different. Let&#39;s understand what is flask and Django.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/836/1*60HibA7pmWH7Q36mSg3CUw.jpeg" /><figcaption><a href="https://miro.medium.com/max/836/1*TbMXbstb039keN-glx2CqQ.jpeg">Image Source</a></figcaption></figure><p>What can be other good examples of tech controversy? Comment below. Would love to know your views.</p><h3><strong>What are Flask and Django?</strong></h3><p>Flask and Django both are the frameworks that are used in creating a web-application. We can create standalone websites with the help of Flask and Django. Flask is classified as a microframework because it does not require particular tools or libraries. It has no database abstraction layer, form validation, or any other components where pre-existing third-party libraries provide common functions.</p><p>But the problem is, which one to choose between the two. We will come to a conclusion by comparing both of its features.</p><p>So let&#39;s talk features,</p><h3>Major Difference</h3><p>The major difference between the two is that Flask is a microframework i.e it does not have any built-in database or auth system. We have to write code from scratch or attach a third-party package to it like <a href="https://flask-login.readthedocs.io/en/latest/">flask-login</a>.</p><p>Whereas, Django has its own <a href="https://en.wikipedia.org/wiki/Object-relational_mapping"><strong>ORM</strong></a><strong>(Object-relational mapping)</strong>, uses data models in the backend, has an inbuilt auth system for its forms. But with flask, you don&#39;t get this privilege.</p><p>Now we can argue on creating a separate data model for the flask. But flask has this basic intention of keeping its package lightweight and rather gives the developers flexibility to use there own data model or package.</p><p>Django is based on the Model-View-Controller framework. The MVC allows the code to be reusable, a lesser line of code, and faster development.</p><h3>Django’s Modular approach</h3><p>In Django, we can divide the whole website into apps. These apps can be used in other projects as well. Writing code in Django is pretty easy.</p><p>Django makes the code easily readable. Its really a hectic work understanding someones else code. Think you have joined a new company, one might spend hours wasting on getting comfortable with the code.</p><p>But when we create an app in Django these files are created automatically.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/275/1*qvuhyMCsIZznOv154KeviQ.png" /></figure><p>Ok. Now even without getting into the code, I can predict that urls.py might be used for routing of URLs to functions. Views.py containing the functions to the related URL endpoint. Models.py be related to something with the database. Easy !!</p><p>A developer can easily create this without wasting much time on thinking about the file names.</p><h3>Talking about security</h3><p>Django has provided inbuilt security feature for these problems:-</p><h4>Cross-site scripting (XSS) protection</h4><p>XSS attacks allow a user to inject client-side scripts into the browsers of other users. This is usually achieved by storing the malicious scripts in the database where it will be retrieved and displayed to other users, or by getting users to click a link which will cause the attacker’s JavaScript to be executed by the user’s browser.</p><p>But using Django’s template we stand against most of these forgeries.</p><h4>Credentials Protection using CSRF</h4><p>CSRF attacks allow a malicious user to execute actions using the credentials of another user without that user’s knowledge or consent.</p><p>Django has built-in protection against most types of CSRF attacks, providing you have <a href="https://docs.djangoproject.com/en/3.1/ref/csrf/#using-csrf">enabled and used it</a> where appropriate.</p><h4>Data Safety (SQL Injection)</h4><p>SQL injection is a type of attack where a malicious user is able to execute arbitrary SQL code on a database. This can result in records being deleted or data leakage.</p><p>Django’s querysets are protected from SQL injection since their queries are constructed using query parameterization.</p><p>So your data remains safe.</p><p>You can read about it in more detail <a href="https://docs.djangoproject.com/en/3.1/topics/security/">here</a>.</p><h3>Admin Dashboard</h3><p>One of the most amazing features of Django is its admin dashboard feature. It provides an in-built Admin dashboard where we can perform CRUD operations. This is an example of the Admin Dashboard Django.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/630/1*16QZbc-0lm--DAqGcpaSZw.png" /><figcaption><a href="https://books.agiliq.com/projects/django-admin-cookbook/en/latest/_images/default_listview.png">Image Source</a></figcaption></figure><p>This UI reduces the load of creating a separate dashboard for admins. The Admins can easily monitor the data from this dashboard.</p><p>During my past years of programming, I really found this feature most interesting. It reduced my workload for building a separate admin dashboard for clients. Most clients are not very versed in programming. So this works as a good to go solution for most cases.</p><p>We can definitely change the UI by using a third-party library like <a href="https://github.com/geex-arts/django-jet">Django-Jet</a>.</p><h3>Flexibility</h3><p>As far as flexibility is concerned, Flask seems to have a complete win over django. Flask is flexible allowing its users to use frameworks according to there needs.</p><h3>Performance</h3><p>Well as far as performance is concerned both deliver pretty good standards with there work.</p><h3>Decetralizatoin of Code</h3><p>When a product grows it becomes absolutely necessary to deliver the content to the users at a faster rate. A user will never stay on your website if the response time is more than a min.</p><p>The solution to this is the decentralization of the code. With Django’s Modularity approach where we have a separate app for every feature, we can simply deploy that app into another <a href="https://en.wikipedia.org/wiki/Virtual_machine">VM </a>(Virtual Machine or <a href="https://en.wikipedia.org/wiki/Virtual_private_server">VPS </a>) and use an API to get faster results. We can do this in the flask as well, required you have followed the modular approach for building your app.</p><p>But yet, due to the lightweight of the flask, I still prefer making APIs with flask.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/747/1*6LgqFWnOe-YrIqqS4Ag6hw.png" /><figcaption><a href="https://avinetworks.com/glossary/microservice/">An example of Mircoservcies also referred as decentralization here.</a></figcaption></figure><h3>Which one to choose?</h3><p>We have only talked about the advantages of Django. Now, let’s take a while to understand flasks usage. As discussed earlier, if you want to have full access over your website its good to go with flask. But this doesn’t promise code reliability, reusability, modular approach. Django is built in a way where we are bound to follow good programming standards.</p><p>I have been into coding for the past few years. Working with startups include building a product from scratch to a level where it matches production standards. There are many things that need to be looked into while building a product. So when thought from the product development point of view, Django is the best solution. Yet if I want to have complete control over my application Flask is the best one.</p><p>In short, if some other package is easing out my work of database, authentication, data-models, initial security. It’s best to stick with it.</p><p>I have been working as a Technology Lead for the past 2 years. And have been into development for a few years.</p><p>I guess frameworks will come and go but what matters is the skillset that you have learned.</p><p>Let&#39;s learn together. If you have read this blog till the end do give a clap and share it with your friends. Let&#39;s connect on LinkedIn. I post active content on <a href="https://www.linkedin.com/in/abhishekpatnaik77/">likedin</a>. If you have any opportunity for me, would love to connect on LinkedIn as well.</p><p><a href="https://www.linkedin.com/in/abhishekpatnaik77/">Abhishek Patnaik - Product Manager - Quantamix Solutions | LinkedIn</a></p><p>Don’t forget to clap, if you made it till the end !!</p><p>Let me know your views in the comment section below. Would love to know your views on it.</p><p>Peace !!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fb9c33ceb4ea" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Extracting Topics with Jointly Embedded Topics]]></title>
            <link>https://abhishekpatnaik77.medium.com/extracting-topics-with-jointly-embedded-topics-291b2eff04ad?source=rss-3b21676c7da2------2</link>
            <guid isPermaLink="false">https://medium.com/p/291b2eff04ad</guid>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[nlp]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[python]]></category>
            <dc:creator><![CDATA[Abhishek Patnaik]]></dc:creator>
            <pubDate>Mon, 31 Aug 2020 10:10:18 GMT</pubDate>
            <atom:updated>2020-08-31T10:10:18.172Z</atom:updated>
            <content:encoded><![CDATA[<p>Extracting keywords with similar meaning from a Text Corpus</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HEHTkm2jOheO8YQ563_3VQ.png" /></figure><p>Ever thought of extracting or searching for similar words from a big text corpus, but then you got thought of the hard work and then sat to code. Because that&#39;s the only option.</p><p>Woooh !!! I got tired of writing the above text.</p><p><a href="https://towardsdatascience.com/topic-modeling-in-nlp-524b4cffeb68">Topic modeling</a> is used to find a topic out of a big text corpus. <a href="https://towardsdatascience.com/lda-topic-modeling-an-explanation-e184c90aadcd">LDA</a> is one of the most common methods of doing it.</p><p>Let&#39;s simplify and understand the problems with old school Latent Dirichlet Allocation and Probabilistic Latent Semantic Analysis. We will also learn about the good-to-go way of extracting keywords. Both use the bag-of-words representation and ignore the semantics.</p><p>Example:- In a <a href="https://www.geeksforgeeks.org/bag-of-words-bow-model-in-nlp/#:~:text=Hence%2C%20Bag%20of%20Words%20model,corresponding%20to%20the%20word%20itself.">BOW</a> representation, India and Indian both would be considered as different words even though they are semantically the same. (If you find this example interesting, give a smile, though I know it&#39;s not.)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/545/1*INYmrG2DK9Eq7d-ZHaXHXg.png" /></figure><p>I have been into NLP for the past 3 years. During these 3 years, I have found the importance of semantics in NLP.</p><p>There have been many such methods that do show promising results on extracting keywords with similar keywords but <a href="https://github.com/ddangelov/Top2Vec">top2vec</a> really gave some promising results. So let&#39;s come to Top2vec way for topic modeling and semantic search.</p><h3><strong>What is Top2vec?</strong></h3><p>In <a href="https://github.com/ddangelov/Top2Vec">top2vec</a>, which leverages joint document and word semantic embedding to find topic vectors.</p><p><a href="https://github.com/ddangelov/Top2Vec">Top2vec</a> is a distributed topic vector which is calculated from dense areas of document vectors. The number of dense areas of documents found in the semantic space is assumed to be the number of prominent topics. The topic vectors are calculated as the centroids of each dense area of document vectors.</p><p>Top2vec is capable of performing the below functionalities after training.</p><ul><li>Get the number of detected topics.</li><li>Get topics.</li><li>Get topic sizes.</li><li>Get hierarchical topics.</li><li>Search topics by keywords.</li><li>Search documents by topic.</li><li>Search documents by keywords.</li><li>Find similar words.</li><li>Find similar documents.</li></ul><p>In the top2vec, topic vector in the semantic embedding represents a prominent topic shared among documents.</p><h3>Model Description:-</h3><p>Let&#39;s discuss the model description of top2vec. The following 4 methods are required to build this model.</p><h4>Step1:-</h4><p><strong>Creating Semantic Embeddings</strong></p><p>This follows the simple concept of placing semantically similar words together in the embedding space and dissimilar words further from each other. Words that represent the best should be present nearest to the dense cloud. Let&#39;s make it easy. Let&#39;s visualize</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/736/1*9wHuE35veipEcJi2Py_4sw.png" /><figcaption>as we see from the figure the words that represent the best are close enough</figcaption></figure><h4>Step 2:-</h4><p><strong>Creating a low dimensional vector space</strong></p><p>Top2vec uses the concepts of UMAP for dimensionality reduction. But why only UMAP and not t-sne or something else. UMAP seems to have better results on large datasets.</p><p>UMAP has several hyper-parameters that determine how it performs dimension reduction. Perhaps the most important parameter is the number of nearest neighbors, which controls the balance between preserving global structure versus local structure in the low dimensional embedding.</p><h4>Step 3:-</h4><p><strong>Finding the dense cloud</strong></p><p>UMAP reduced vectors are then used to find the clouds with the highly-dense documents that have similar documents. But there&#39;s a problem here, i.e, the data will be sparse at times, full of noises, and no prominent indication of topics.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/533/1*5gAY-eUMjkoyriH5heaCUg.png" /></figure><p>HDBSCAN is used to find the dense areas of document vectors, as it was designed to handle both noise and variable density clusters. It solves the issue of sparse density. HDBSCAN assigns a label to each dense cluster of document vectors and assigns a noise label to all document vectors that are not in a dense cluster. Which makes it more effective.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/522/1*4NhhhrRMq220cl0FdaLwdw.png" /><figcaption>Red areas are the noises and the other colors are the labels from HDBSCAN</figcaption></figure><h4>Step 4:-</h4><p><strong>Calculate Centroids in Original Dimensional Space</strong></p><p>Now, as we have the labels for each cluster which was derived from Step3, we can calculate the topic vectors. The methods for detecting centroids are by the arithmetic mean of all the document vectors in the same dense cluster or geometric mean or using probabilities from the confidence of clusters.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/613/1*mhCtK21m_e8Ip0KlwH_-jQ.png" /><figcaption>Red outliers are not used to calculate centroid.</figcaption></figure><p>Now let&#39;s test it out on code.</p><h4>Install top2vec</h4><pre>pip install top2vec</pre><p>Training it on a Sklearn News Dataset:-</p><pre>from top2vec import Top2Vec<br>from sklearn.datasets import fetch_20newsgroups<br><br>newsgroups = fetch_20newsgroups(subset=&#39;all&#39;, remove=(&#39;headers&#39;, &#39;footers&#39;, &#39;quotes&#39;))<br><br>model = Top2Vec(documents=newsgroups.data, speed=&quot;learn&quot;, workers=8)</pre><p>Below is a result of a used case scenario of searching topics in the word corpus.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2B5bdmBE4JueFmpUCu8H0Q.png" /></figure><p>If you like this do clap, show your love by sharing. You can follow me on medium for more such content.</p><p><strong>Top2vec Github</strong>: <a href="https://github.com/ddangelov/Top2Vec">https://github.com/ddangelov/Top2Vec</a></p><p>Let&#39;s connect on Linkedin.</p><p><a href="https://www.linkedin.com/in/abhishekpatnaik77/">Abhishek Patnaik - Product Manager - Quantamix Solutions | LinkedIn</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=291b2eff04ad" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>