<?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[Winkl Insights - Medium]]></title>
        <description><![CDATA[Our learnings on team building, product, growth and sales as we build Winkl. - Medium]]></description>
        <link>https://medium.com/winkl-insights?source=rss----f9a703a784dc---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>Winkl Insights - Medium</title>
            <link>https://medium.com/winkl-insights?source=rss----f9a703a784dc---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 03:18:41 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/winkl-insights" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[0% attrition till date!]]></title>
            <link>https://medium.com/winkl-insights/0-attrition-till-date-426400b38ce4?source=rss----f9a703a784dc---4</link>
            <guid isPermaLink="false">https://medium.com/p/426400b38ce4</guid>
            <category><![CDATA[culture]]></category>
            <category><![CDATA[team]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[team-building]]></category>
            <category><![CDATA[hiring]]></category>
            <dc:creator><![CDATA[Thakur Rahul Singh]]></dc:creator>
            <pubDate>Tue, 26 Jan 2021 08:14:05 GMT</pubDate>
            <atom:updated>2021-01-26T08:14:05.181Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lqOc0bNJHLj3YpO98BDnPA.jpeg" /><figcaption>We are not building Winkl like a family, for us it’s like a professional sports team</figcaption></figure><blockquote>The Winkl journey over the last few years has taken us through extremes — lows to highs and to lows again, very quickly. While there were many milestones accomplished and things built, we made our fair share of mistakes. As a team, we are very proud of where we’ve reached in a seemingly very competitive industry, but if there’s one thing that we are really proud of, it is the culture we’ve managed to build at Winkl.</blockquote><p>Interestingly, neither Nikhil or I have any professional work experience. We started Winkl right out of college and didn’t have the opportunity to experience the work culture of any other company. The little experience I had was when I was interning at Flipkart for a few months, that’s about it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*u4ep9v59kuMe3BR-" /><figcaption>Our first real team, back in June 2018</figcaption></figure><p>At the time we started, we never thought about it much, but in hindsight, it does have its fair share of advantages and disadvantages.</p><p><strong>A disadvantage</strong> — when you consider the fact that we did not have any first hand experience when it come to observing how teams function and a playbook to follow. Often, when people quit working at a corporate/startup to start their own, they have a reference point to begin with. We did not have that. <strong>And an advantage</strong>, as it was like a blank canvas for us, to try and create a culture from scratch without knowing too much about how other companies get built.</p><p>We are a team of 25 now, and when we looked back at our journey over the last 3 years, we realized that not a single team-mate of ours has voluntarily quit Winkl. It may not be the same in the future as some might quit to start their own startups (a good sign) or for reasons beyond work, at this juncture it feels like a major accomplishment.</p><p>This post is essentially how we have tried to build the culture at Winkl. We’re far from perfect and have many flaws, the hope is to try and help other startups with pointers on what I think worked for us.</p><h4>We built a place that we would love working at as employees</h4><p>We often come across many horror stories of our friends working at other companies on various policies &amp; rules that just didn’t feel right. It almost looked like the companies were treating their employees as school kids with all sorts of restrictions and policies in place that get enforced to shape behavior. While this might be necessary to ensure productivity for the 10% employees who may not be contributing, it spoils the experience for the remaining 90%, which is a bad tradeoff.</p><blockquote>We tried and created a culture where we would love to work as employees. A place where high quality work is valued more than hours worked. We didn’t care about the number of holidays/offs someone took as long as they were responsible and took ownership.</blockquote><h4>Hiring people with similar outlook towards work</h4><p>We hired people who are very similar to us in terms of their values and outlook towards work. This helped us a lot as they then hired their own team with similar values and the non performers automatically get highlighted as they do not fit in very well.</p><p>An example — the moment someone brings up the number of hours they have worked as a metric to show their contribution, its a red flag and we know he/she may not be the right fit for us. Another one is when we end up micromanaging someone, its a red flag as micromanaging is something we do not like to do as a company.</p><p><strong>Some of the other basic red flags —</strong></p><ol><li>Someone you cannot hangout with</li><li>If they get offended very easily</li><li>Expectation of being spoon fed</li><li>If they agree with you on everything without questioning</li><li>Lack of curiosity, ethics and strong values &amp; opinions.</li><li>Bias for action without thought.</li><li>If they come from highly affluent families.</li></ol><p>These red flags may not work for all and you will have to devise your own way of filtering for these, but this is a work in progress. While these work as basic filters, they are not enough. We try and look for something that really stands out about them, either in professional or personal life.</p><h4>We do not micro-manage</h4><p>We’ve noticed that the best people do not like being micro managed. You do not want to check-in with them every few hours on what they are up to. Usually, this ends up happening when you are not sure if they are contributing meaningfully and want to check-in with them for your personal satisfaction. But the fact of the matter is, if you are forced to micromanage someone, its not a great sign, it probably means that they are waiting for you to reach out to them and offer help, means they are not being proactive and that’s a big red flag.</p><p>It’s best to part ways with such people as they end up draining your energy and not contributing much to the overall progress.</p><h4>We do not give fancy perks, we optimize for freedom and flexibility</h4><p>Often, even at some of the top startups we know of, fancy offices, ping pong tables, free lunches, team outings etc., are disguised as culture. While those are great to have, they don’t define culture at all.</p><p>Culture, is the little things that define how the team functions, what they value and what they ignore on a day to day basis.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*IqFV_zZ-Proe2jIy-gel7g.gif" /><figcaption>The Bangalore team’s recent trek</figcaption></figure><p><strong>Freedom &amp; flexibility—</strong> we think its very important for our team to make their own decisions with most things concerning their work. The advantage of this is, when people decide for themselves, they automatically become more accountable and start taking ownership. We do not tell them when to start their work or when to take an off. If they think something is worth trying out, they are free to do so. There are many such examples but the larger point is, treat them like mature adults who know what they are doing.</p><p>We simply help the team with clarity on the larger goals and objectives with clear focus on who is responsible for what.</p><h4>We put people in uncomfortable positions very quickly</h4><p>This is analogous to someone jumping off a cliff and assembling an airplane on their way down. Once we see good potential in someone, we quickly promote them to lead teams and give them a lot of responsibility on their shoulders.</p><p><strong>This usually has two outcomes, people succumb under the pressure or thrive and shine. </strong>In the former case, we simply take off some responsibility and let the person contribute individually. In the latter case, we have found ourselves a leader, which is very valuable. In fact, all our team leads have gone through this exact process and have thrived.</p><p>These are just some of the things that have worked for us and do not do justice to the entire process of culture building. There’s a lot more to do, a lot to build and miles to go.</p><p>Now, back to work!</p><p><em>At Winkl, we are pioneering influencer &amp; creator collaborations for brands and agencies. </em><a href="http://winkl.co/"><em>Learn more about what we do.</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=426400b38ce4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/winkl-insights/0-attrition-till-date-426400b38ce4">0% attrition till date!</a> was originally published in <a href="https://medium.com/winkl-insights">Winkl Insights</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 future of collaborations in a $400bn/year industry: Winkl]]></title>
            <link>https://medium.com/winkl-insights/building-the-future-of-collaborations-in-a-400bn-year-industry-winkl-940cbc319c32?source=rss----f9a703a784dc---4</link>
            <guid isPermaLink="false">https://medium.com/p/940cbc319c32</guid>
            <category><![CDATA[collaboration]]></category>
            <category><![CDATA[content]]></category>
            <category><![CDATA[winkl]]></category>
            <dc:creator><![CDATA[Thakur Rahul Singh]]></dc:creator>
            <pubDate>Fri, 02 Oct 2020 11:29:54 GMT</pubDate>
            <atom:updated>2020-10-02T13:52:53.635Z</atom:updated>
            <content:encoded><![CDATA[<p>One of the odd and exciting things about building a startup is that you and your team get to define the ambition and scale at which you want to operate. Leaving aside a few hygienic things such as the ability to execute, speed, etc., the ambition and its scale is what sets apart successful teams from the rest.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yYGXLRXpHynUx2RzOGvXAQ.jpeg" /></figure><p>At Winkl, one of the things we wanted to do from the start was to build something that creates an impact at scale. However, for a long time, it was not fully clear as to what this ‘scale’ really meant or how it would manifest. But today, after spending more than 2.5 years building Winkl, we think there’s some clarity.</p><p>For those of you who don’t know, <a href="http://winkl.co/">Winkl</a> operates in the influencer marketing space where we help brands &amp; agencies execute influencer marketing campaigns at scale using data, technology-driven automation, and data-backed strategy. In simple words, it&#39;s a <a href="https://www.loom.com/share/0b46c9eade6449948cb7f3ccda8f7275">web app</a> that helps our customers (brands &amp; agencies) run influencer marketing campaigns, end to end.</p><p><strong>Is this our end game?</strong></p><h4>The larger picture</h4><p>There’s a strong shift that the $400bn/year content industry is witnessing today. It is the unbundling of media brands and ad/creative agencies into smaller groups and individuals.</p><blockquote>We believe that a significant part of the world’s creative output in the future will come from smaller/niche creative groups or individual creators vs large media houses and agencies.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/770/0*mB6eQu8TCi94lOx5.png" /></figure><p>This means, there is going to be a lot more business happening at the long tail than ever before.</p><h4>What does this mean?</h4><p>This change means that working with creators is going to get a lot more exciting and creative with a huge variety in ideas, this will also result in difficulty for brands with respect to identifying and working with the right creators, the processes, communication, standardization, and transparency.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FIIj8QhRfdnY%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DIIj8QhRfdnY&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FIIj8QhRfdnY%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/fa250cca661f5b3ea93a8e68d53a6d9c/href">https://medium.com/media/fa250cca661f5b3ea93a8e68d53a6d9c/href</a></iframe><p><strong>This is where Winkl comes in. </strong>Our aspiration is to enable creators to find relevant projects and campaigns to work on which they find exciting.</p><p>And at the same time help brands to discover the most promising creators for their use case while taking care of issues such as creator quality and vetting, briefs and communication, onboarding, commercials, negotiations, and contractual obligations while maintaining a high amount of trust and transparency between all the parties.</p><h4>Why does it matter?</h4><p>Simple words, content is everywhere. Right from the recipe video you watched to cook up your favorite dish during the lockdown or the podcast you’ve been listening to on your evening runs. The recent advertisement created by Cred featuring Bappi Da, Anil Kapoor and Madhuri Dixit to the Whatsapp forward your colleague shared with you. Content is nothing but a manifestation of language and communication. The point is, it&#39;s everywhere.</p><p>As the technology evolved and became more accessible, it had a direct impact on content creation and its democratization. In essence, you and I are creators too even if we clicked a simple picture.</p><p>This suddenly led to an exponential growth in the creative output of the world, and some of these talented creators, thanks to distribution platforms like Facebook, YouTube, Instagram, TikTok, etc., achieved popularity.</p><p>Their talent was now being noticed by brands and others who wanted to collaborate with them.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FSjhBeIz1tDI%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DSjhBeIz1tDI&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FSjhBeIz1tDI%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="640" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/3f0e5c56a3d9a2a79c0fd45a8a2252e2/href">https://medium.com/media/3f0e5c56a3d9a2a79c0fd45a8a2252e2/href</a></iframe><p>A recent example is Yashraj Mukhate who started receiving a flood of collaboration requests after his music video <a href="https://www.youtube.com/watch?v=SjhBeIz1tDI&amp;ab_channel=YashrajMukhate">“Rasode Mein Kaun Tha”</a> went viral.</p><h4>Why now?</h4><p>Apart from the points earlier, another interesting shift is that brands are now more open to working with individual creators instead of sticking with the status quo and working with reputed firms and media/production houses.</p><p>When all this comes together, magic happens! Brands get to work with creative talent that they never had access to before. This could result in collaborations which create some great outcomes for brands and offer interesting projects for creators to work on.</p><p>That’s something.</p><p>When we look at the potential of the content industry and what’s possible with technology, it&#39;s a tingling feeling :)</p><p>Now, back to work!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=940cbc319c32" width="1" height="1" alt=""><hr><p><a href="https://medium.com/winkl-insights/building-the-future-of-collaborations-in-a-400bn-year-industry-winkl-940cbc319c32">Building the future of collaborations in a $400bn/year industry: Winkl</a> was originally published in <a href="https://medium.com/winkl-insights">Winkl Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How paying our creators upfront improved the output and reduced our work]]></title>
            <link>https://medium.com/winkl-insights/how-paying-our-creators-upfront-improved-the-output-and-reduced-our-work-a848d2f29e0b?source=rss----f9a703a784dc---4</link>
            <guid isPermaLink="false">https://medium.com/p/a848d2f29e0b</guid>
            <category><![CDATA[startup-lessons]]></category>
            <category><![CDATA[winkl]]></category>
            <category><![CDATA[lessons-learned]]></category>
            <category><![CDATA[influencer-marketing]]></category>
            <category><![CDATA[startup-life]]></category>
            <dc:creator><![CDATA[Thakur Rahul Singh]]></dc:creator>
            <pubDate>Fri, 05 Jul 2019 16:09:19 GMT</pubDate>
            <atom:updated>2019-07-05T16:09:19.255Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*S24gAHbnsc6N7y1KM_Kyxg.jpeg" /></figure><p>As we built out the <a href="https://winkl.co/">Winkl</a> app over a few months and were going into testing, one of the requirements we had was to seed the app with content so that when the testers arrived, the app wasn’t a ghost-town and there was some pre existing content, which would engage the testers and also set a benchmark as to what content to create on the platform.</p><blockquote><em>At Winkl, we’re building a community of creators, bloggers &amp; influencers and help them create their media kits, get more detailed insights into their influence and learn &amp; share on how to create great content.</em></blockquote><p>For this purpose, we reached out to some of the good creators from our community and asked them to help us out with this. In exchange, we offered them a small token amount as gratitude. The arrangement was simple. They would create a few posts per week and we would pay them at the end of 4 weeks.</p><p><strong>Here’s where the fun begins!</strong></p><p>We were supposed to get the first set of videos in a week. We waited patiently. And then some more. It was almost 15 days now and we hadn’t got a single one, even after multiple followups. I was wondering on what would get these creators to give us the content quickly and then something struck.</p><blockquote>What if we pay them in advance for a month and see what happens?</blockquote><p>The idea was that if we trust them blindly upfront, they would feel the urgency to send us the posts quickly, out of accountability and guilt. Obviously, there were questions around what if they take our money and don’t respond? What if they respond but still take their own time? We decided to keep the questions aside for a bit and go ahead with the plan.</p><p>My colleague <a href="https://www.instagram.com/rawatsurabhi/">Surabhi</a> transferred the payment to them upfront and waited. Within two days, we started receiving the videos we wanted.<strong> </strong>Within a week, we had all that we wanted with hardly a followup or two.</p><p><strong>The idea worked! </strong>The learning —</p><blockquote>When you trust people blindly, they feel accountable and deliver.</blockquote><p>This is something I have relied on, with our team as well. Trust them blindly and they will deliver. Of course, sometimes you have bad actors/instances as well, but the pros of trusting everyone blindly outweighs the cons.</p><blockquote><strong>Follow </strong><a href="https://medium.com/winkl-insights"><strong>Winkl Insights</strong></a><strong> on Medium as I go about sharing our learnings on how we’re building an influencer community platform from India, for the world.</strong></blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a848d2f29e0b" width="1" height="1" alt=""><hr><p><a href="https://medium.com/winkl-insights/how-paying-our-creators-upfront-improved-the-output-and-reduced-our-work-a848d2f29e0b">How paying our creators upfront improved the output and reduced our work</a> was originally published in <a href="https://medium.com/winkl-insights">Winkl Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Setting up a safe hashed password system to prevent internal/external access.]]></title>
            <link>https://medium.com/winkl-insights/setting-up-a-safe-hashed-password-system-to-prevent-internal-external-access-932f63dc59d5?source=rss----f9a703a784dc---4</link>
            <guid isPermaLink="false">https://medium.com/p/932f63dc59d5</guid>
            <category><![CDATA[api]]></category>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[passwords]]></category>
            <category><![CDATA[security]]></category>
            <category><![CDATA[ruby-on-rails]]></category>
            <dc:creator><![CDATA[Nikhil Kumar]]></dc:creator>
            <pubDate>Sun, 21 Apr 2019 13:36:55 GMT</pubDate>
            <atom:updated>2019-04-21T13:36:55.547Z</atom:updated>
            <content:encoded><![CDATA[<h3>Setting up a hashed password system to prevent internal/external access.</h3><p>This is a simple article which will explain how you can go about setting up a simple system to prevent passwords from being accessible to any one. The motive of this post is to show that using rails, it is extremely simple to setup.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HTTbaSF-muw6i7CcToVPzQ.jpeg" /></figure><p>I am planning to create a series of posts built upon this post that picks one topic and helps implement it easily, using rails. Let me know what you think. Lets get started:</p><p>Here is a <a href="https://github.com/nikhil4092/RailsAPITutorial">Github</a> link to the full code.</p><p>Start off by creating a <strong>new rails API app</strong> with this simple single line command:</p><pre>rails new new_app --api</pre><p>Install all gems by running:</p><pre>gem install bundle</pre><pre>bundle</pre><p><strong>Create a user scaffold:</strong></p><pre>rails g scaffold user name, email, profile_picture</pre><p>Your file /db/migrate/*_create_users.rb looks like this:</p><pre>class CreateUsers &lt; ActiveRecord::Migration[5.2]<br>  def change<br>    create_table :users do |t|<br>      t.string :name<br>      t.string :email, index: true<br>      t.string :profile_picture<br>      t.timestamps<br>    end<br>  end<br>end</pre><p>Your file /app/models/user.rb looks like this:</p><pre>class User &lt; ApplicationRecord<br>    validates :name, presence: true<br>    validates :email, presence: true, uniqueness: true, on: :create<br>end</pre><pre>//validates helps you place conditions on certain fields</pre><p>Your file /app/controllers/users_controller.rb looks like this:</p><pre>class UsersController &lt; ApplicationController<br>  before_action :set_user, only: [:show, :update, :destroy]</pre><pre># GET /users<br>  def index<br>    @users = User.all</pre><pre>render json: @users<br>  end</pre><pre># GET /users/1<br>  def show<br>    render json: @user<br>  end</pre><pre># POST /users<br>  def create<br>    @user = User.new(user_params)</pre><pre>if @user.save<br>      render json: @user, status: :created, location: @user<br>    else<br>      render json: @user.errors, status: :unprocessable_entity<br>    end<br>  end</pre><pre># PATCH/PUT /users/1<br>  def update<br>    if @user.update(user_params)<br>      render json: @user<br>    else<br>      render json: @user.errors, status: :unprocessable_entity<br>    end<br>  end</pre><pre># DELETE /users/1<br>  def destroy<br>    # @user.destroy //prevent delete action ATM<br>    render json: {action: false}<br>  end</pre><pre>private<br>    # Use callbacks to share common setup or constraints between actions.<br>    def set_user<br>      @user = User.find(params[:id])<br>    end</pre><pre># Only allow a trusted parameter &quot;white list&quot; through.<br>    def user_params<br>      params.require(:user).permit(:name, :email, :profile_picture)<br>    end<br>end</pre><p>Create and migrate your database:</p><pre>rails db:create</pre><pre>rails db:migrate</pre><p>This is all you need for CRUD(Without the D) for a user. We will now create a separate password table to store user passwords. I prefer this method as it helps easily return user objects without the worry of including passwords in the object by mistake. Here is an example:</p><pre>Dummy function:</pre><pre>def get_user<br>   user = User.first //Here I can pick the user without passwords<br>   render json {user: user} <br>end</pre><p><strong>Create a user password table:</strong></p><pre>rails g model UserPassword</pre><p>Your file /db/migrate/*_create_user_passwords.rb looks like this:</p><pre>class CreateUserPasswords &lt; ActiveRecord::Migration[5.2]<br>  def change<br>    create_table :user_passwords do |t|<br>      t.belongs_to :user, index: true<br>      t.string :password_hash<br>      t.timestamps<br>    end<br>  end<br>end</pre><p>Your file /app/models/user_password.rb looks like this:</p><pre>class UserPassword &lt; ApplicationRecord<br>    belongs_to :user<br>end</pre><p>Update your file /app/models/user.rb to look like this:</p><pre>class User &lt; ApplicationRecord<br>    has_one :user_password, dependent: :delete<br>    validates :name, presence: true<br>    validates :email, presence: true, uniqueness: true, on: :create<br>end</pre><pre>//validates helps you place conditions on certain fields</pre><p>Migrate your user_passwords table:</p><pre>rails db:migrate</pre><p>Adding BCRYPT to Gemfile. Bcrypt is used to hash passwords so that they cannot be un-hashed to reveal the password.</p><p>It looks like this:</p><pre>source &#39;<a href="https://rubygems.org&/#39;">https://rubygems.org&#39;</a><br>git_source(:github) { |repo| &quot;<a href="https://github.com/#{repo}.git">https://github.com/#{repo}.git</a>&quot; }</pre><pre>ruby &#39;2.3.4&#39;</pre><pre># Bundle edge Rails instead: gem &#39;rails&#39;, github: &#39;rails/rails&#39;<br>gem &#39;rails&#39;, &#39;~&gt; 5.2.3&#39;<br># Use sqlite3 as the database for Active Record<br>gem &#39;sqlite3&#39;<br># Use Puma as the app server<br>gem &#39;puma&#39;, &#39;~&gt; 3.11&#39;<br># Build JSON APIs with ease. Read more: <a href="https://github.com/rails/jbuilder">https://github.com/rails/jbuilder</a><br># gem &#39;jbuilder&#39;, &#39;~&gt; 2.5&#39;<br># Use Redis adapter to run Action Cable in production<br># gem &#39;redis&#39;, &#39;~&gt; 4.0&#39;<br># Use ActiveModel has_secure_password<br># gem &#39;bcrypt&#39;, &#39;~&gt; 3.1.7&#39;</pre><pre># Use ActiveStorage variant<br># gem &#39;mini_magick&#39;, &#39;~&gt; 4.8&#39;</pre><pre># Use Capistrano for deployment<br># gem &#39;capistrano-rails&#39;, group: :development</pre><pre># Reduces boot times through caching; required in config/boot.rb<br>gem &#39;bootsnap&#39;, &#39;&gt;= 1.1.0&#39;, require: false</pre><pre><strong>gem &#39;bcrypt-ruby&#39;, :require =&gt; &#39;bcrypt&#39;</strong></pre><pre># Use Rack CORS for handling Cross-Origin Resource Sharing (CORS), making cross-origin AJAX possible<br># gem &#39;rack-cors&#39;</pre><pre>group :development, :test do<br>  # Call &#39;byebug&#39; anywhere in the code to stop execution and get a debugger console<br>  gem &#39;byebug&#39;, platforms: [:mri, :mingw, :x64_mingw]<br>end</pre><pre>group :development do<br>  gem &#39;listen&#39;, &#39;&gt;= 3.0.5&#39;, &#39;&lt; 3.2&#39;<br>  # Spring speeds up development by keeping your application running in the background. Read more: <a href="https://github.com/rails/spring">https://github.com/rails/spring</a><br>  gem &#39;spring&#39;<br>  gem &#39;spring-watcher-listen&#39;, &#39;~&gt; 2.0.0&#39;<br>end</pre><pre># Windows does not include zoneinfo files, so bundle the tzinfo-data gem<br>gem &#39;tzinfo-data&#39;, platforms: [:mingw, :mswin, :x64_mingw, :jruby]</pre><p>Update your gems:</p><pre>bundle</pre><p>You will need two specific methods from Bcrypt to generate secure passwords:</p><p>The <strong>create</strong> function is used to generate a hashed password to be used. This password cannot be un-hashed easily.</p><pre>hashed_password = <em>BCrypt</em>::<em>Password</em>.create(&quot;Password&quot;, cost: 13) //12 is default</pre><pre>//The cost determines how hard it is to un-hash your password. The higher, the more hard. Creating passwords with higher cost take more time.</pre><p>Now all you need to do is modify your user_controller.rb file to take have login and create APIs to access user data.</p><p><strong>Create API:</strong> Creates a user and adds their password as a hashed string which cannot be un-hashed.</p><pre>  def create<br>    @user = User.new(user_params)<br>    if !params[&#39;password&#39;].nil?<br>      password = BCrypt::Password.create(params[&#39;password&#39;])<br>      if @user.save<br>        up = UserPassword.create(user_id: @user.id, password_hash: password)<br>        render json: @user, status: :created, location: @user<br>      else<br>        render json: @user.errors, status: :unprocessable_entity<br>      end<br>    else<br>      render json: {status: &#39;no password&#39;}<br>    end<br>  end</pre><p><strong>Update API:</strong> Updates user details based on match of password. The password sent is hashed using a similar process during creation and is compared with the existing hashed password. This will help find if its the same password but we will not know what the exact password is till you compare the password. This makes it way harder to find what the password is.</p><pre>  def update<br>    @db_password = BCrypt::Password.new(@user.user_password.password_hash)<br>    if @db_password == params[&#39;password&#39;]<br>      if @user.update(user_params)<br>        render json: @user<br>      else<br>        render json: @user.errors, status: :unprocessable_entity<br>      end<br>    else<br>      render json: {status: &#39;password missmatch&#39;}<br>    end<br>  end</pre><p><strong>Login API: </strong>Picks the user from the email ID sent and compares the password set for that user with the one that is sent during this request by hashing it and comparing it. Similar to how it was done during update. User object is returned on successful match.</p><pre>  def login<br>    @existing_user = User.find_by_email(params[&#39;email&#39;])<br>    if @existing_user<br>      @db_password = BCrypt::Password.new(@existing_user.user_password.password_hash)<br>      if @db_password == params[&#39;password&#39;]<br>        render json: @existing_user<br>      else<br>        render json: {status: &#39;exists&#39;}<br>        return<br>      end<br>    else<br>      render json: {status: false}<br>      return<br>    end<br>  end</pre><p>Your /app/controllers/user_controller.rb will look like this at the end of it:</p><pre>class UsersController &lt; ApplicationController<br>  include BCrypt<br>  # POST /users<br>  def create<br>    @user = User.new(user_params)<br>    if !params[&#39;password&#39;].nil?<br>      password = Password.create(params[&#39;password&#39;])<br>      if @user.save<br>        up = UserPassword.create(user_id: @user.id, password_hash: password)<br>        render json: @user, status: :created, location: @user<br>      else<br>        render json: @user.errors, status: :unprocessable_entity<br>      end<br>    else<br>      render json: {status: &#39;no password&#39;}<br>    end<br>  end</pre><pre># PATCH/PUT /users/1<br>  def update<br>    @db_password = Password.new(@user.user_password.password_hash)<br>    if @db_password == params[&#39;password&#39;]<br>      if @user.update(user_params)<br>        render json: @user<br>      else<br>        render json: @user.errors, status: :unprocessable_entity<br>      end<br>    else<br>      render json: {status: &#39;password missmatch&#39;}<br>    end<br>  end</pre><pre>def login<br>    @existing_user = User.find_by_email(params[&#39;email&#39;])<br>    if @existing_user<br>      @db_password = Password.new(@existing_user.user_password.password_hash)<br>      if @db_password == params[&#39;password&#39;]<br>        render json: @existing_user<br>      else<br>        render json: {status: &#39;exists&#39;}<br>        return<br>      end<br>    else<br>      render json: {status: false}<br>      return<br>    end<br>  end</pre><pre># DELETE /users/1<br>  def destroy<br>    # @user.destroy<br>    render json: {action: false}<br>  end</pre><pre>private<br>    # Use callbacks to share common setup or constraints between actions.<br>    def set_user<br>      @user = User.find(params[:id])<br>    end</pre><pre># Only allow a trusted parameter &quot;white list&quot; through.<br>    def user_params<br>      params.require(:user).permit(:name, :email, :profile_picture)<br>    end<br>end</pre><p>You will need to make some small changes with your /config/routes.rb:</p><pre>Rails.application.routes.draw do<br>  resources :users<br>  get &#39;/login&#39;, to: &#39;users#login&#39;<br>end</pre><p>Testing the app:</p><pre>rails s</pre><p><strong>Create User:</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/742/1*UrhLb81UoYm_ilMudpkyAQ.png" /></figure><p><strong>User Login:</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/833/1*FLPBq7WupmzJeIFwzlV0AQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/833/1*HKcobhlEKCbILdUqXErgTA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/831/1*LftkxT5XwNMIjy0P2bKJ8Q.png" /></figure><p><strong>Hashed Password:</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/958/1*lRt2v_j19ZiDQogkWQtpyQ.png" /></figure><p>Thanks for reading through. Please feel free to leave your feedback and clap if you found this useful. Subscribe to get more posts like this. Any questions? Please leave it in the comments section below. Will try to be as helpful as possible. You can also follow me on <a href="https://twitter.com/n1khil_kumar">Twitter</a>.</p><p><a href="https://github.com/nikhil4092/RailsAPITutorial">nikhil4092/RailsAPITutorial</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=932f63dc59d5" width="1" height="1" alt=""><hr><p><a href="https://medium.com/winkl-insights/setting-up-a-safe-hashed-password-system-to-prevent-internal-external-access-932f63dc59d5">Setting up a safe hashed password system to prevent internal/external access.</a> was originally published in <a href="https://medium.com/winkl-insights">Winkl Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[FACE RECOGNITION: How to deal with people that were not part of training data]]></title>
            <link>https://medium.com/winkl-insights/face-recognition-how-to-deal-with-people-that-were-not-part-of-training-data-36fab27faabc?source=rss----f9a703a784dc---4</link>
            <guid isPermaLink="false">https://medium.com/p/36fab27faabc</guid>
            <category><![CDATA[face-recognition-software]]></category>
            <category><![CDATA[convolution-neural-net]]></category>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[face-recognition]]></category>
            <dc:creator><![CDATA[Chhitij Saraswat]]></dc:creator>
            <pubDate>Sun, 21 Apr 2019 13:35:37 GMT</pubDate>
            <atom:updated>2019-04-21T13:35:37.329Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Hmqbvj-Y5jCkz5-53o8G_g.jpeg" /></figure><p>There are a lot of articles on FaceNet but hardly any discuss the issues faced in developing an actual product using it. Achieving high accuracy such as over 99% is demonstrated. This accuracy hides the fact that the tests are based on images of the same people that were used for training the model. In real life scenarios, this is never the case. The model tries to predict every unknown face as one of the classes in the training data, even the ones that don’t belong to any. This problem occurs when we intend to recognise a few people and ignore the others.</p><p>One possible solution for this can be to set a high confidence value as the threshold for predicting a class for a face but it might come at a cost of losing out on recognition of actual faces in some cases.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*7hOjHp24dTfLdS744qcIEQ.jpeg" /></figure><p>This article explains how to generalise a face recognition system for almost any face and increase the overall accuracy of your model along the way. To understand the whole process of Face Recognition using FaceNet can be found <a href="https://medium.com/@ageitgey/machine-learning-is-fun-part-4-modern-face-recognition-with-deep-learning-c3cffc121d78">here</a>. The basic code used for developing this system was taken from <a href="https://github.com/krasserm/face-recognition">here</a>.</p><p>To develop a system that must recognise only a set of people who are authorised, such as in offices, and ignore the others, we need make a few changes. If we add a new class of images to our training data that tries to match every unknown face, we can differentiate among people we want to be identified by the model and the ones not to be recognised. Now the question arises as to what kind of data would be suitable for this case.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/577/1*_szmsexsY_GLf_ui_5XaMw.jpeg" /></figure><p>Relax, it’s not that hard but first, lets understand how triplet loss works. The triplet loss algorithm tries to find features that best differentiate among faces in training set. It takes a set of triplets, one anchor image, a positive example and a negative example, and generates embeddings for each image.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1016/1*H-gKds4bIXuGs7TvuusY5w.png" /></figure><p>The goal of the triplet loss is to make sure that:</p><ul><li>Two examples with the same label have their embeddings close together in the embedding space</li><li>Two examples with different labels have their embeddings far away.</li></ul><p>The model is bound to learn more minute and exact features if it is trained to differentiate between two similar looking faces. The probability of training on similar faces increases with the more number of faces we have in our training data. Since we only need to predict for a certain set of classes, we can focus on making the model learn the features for these set of classes. Why this works?</p><ol><li>Consider two very similar looking people A, a person we need to recognise, and B, a person not to be recognised.</li><li>Since we have A in our set of people that need to be recognised by our system and B must not be recognised, we can have A in one class and B can be in any other classes.</li><li>For the above scenario, triplet loss will differentiate between A and B, resulting in learning better features for A (not talking about B as for us it doesn’t matter).</li><li>Since B is in another class, if B actually walks into the camera, he would not be recognised as A as it would have been the case if B was not in our data set.</li></ol><p>The next concern would be, where to put this image of B? Since we can’t have a new class for every B that we have for every A, we can make a new class that contains only faces that don’t need to be recognised. So, we have an exclusive class that only contains Bs for all As. This still doesn’t solve our problem. How do we come up with B, a person who looks similar, for every A?</p><p>As said earlier, the probability of training on similar faces increases by increasing total number of people in the data set. One way to get images of people is a simple google search. We can use a chrome extension, <a href="https://chrome.google.com/webstore/detail/fatkun-batch-download-ima/nnjjahlikiabnchcpehcpkdeckfgnohf?hl=en">Fatkun Batch Download Image</a> for downloading all images for this search result, on that particular page. This gives us a variety of faces and helps optimise our model to learn better features. Moreover, every person who is not recognised would be classified to this class.</p><p>RESULTS:</p><p>Using the above approach, the accuracy increased from 96% to 99% using the same set of data (10 classes with 10 images each), except this new class of various people (559 images), besides also being able to differentiate people who are not supposed to be recognised. The actual accuracy of any ML model depends on the data that it was trained on, hence actual values might differ for every data set.</p><p>The code for this can be found at <a href="https://github.com/Chhitij07/face-recognition">https://github.com/Chhitij07/face-recognition</a>. The code still needs to be cleaned to improve performance but for right now, it still works. Steps:</p><ol><li>Create a folder for each class in images folder, including a class for the set of images from google search. If you want to capture images for a new class, it can be done by running add_training_data.py</li><li>The dependencies for the code can be found in the requirements.txt file.</li><li>Run the following command:</li></ol><pre>python face_train.py</pre><p>You only need to do this once. Before terminating, it will start running the webcam to check the results.</p><p>Once the model is trained, it can be used by running the command :</p><pre>python face_recognize.py</pre><p>This file still needs to be improved as it takes some time executing some redundant code. The image used to predict is stored at test/random/</p><p>Feedback would be appreciated.</p><p>I work at <a href="http://winkl.co/">Winkl</a> and building some cool stuff. Feel free to reach out to me :)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=36fab27faabc" width="1" height="1" alt=""><hr><p><a href="https://medium.com/winkl-insights/face-recognition-how-to-deal-with-people-that-were-not-part-of-training-data-36fab27faabc">FACE RECOGNITION: How to deal with people that were not part of training data</a> was originally published in <a href="https://medium.com/winkl-insights">Winkl Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The influencer wave is coming!]]></title>
            <link>https://medium.com/winkl-insights/the-influencer-wave-is-coming-ad83e61bebe0?source=rss----f9a703a784dc---4</link>
            <guid isPermaLink="false">https://medium.com/p/ad83e61bebe0</guid>
            <category><![CDATA[influencers]]></category>
            <category><![CDATA[brands]]></category>
            <category><![CDATA[influencer-marketing]]></category>
            <category><![CDATA[winkl]]></category>
            <dc:creator><![CDATA[Thakur Rahul Singh]]></dc:creator>
            <pubDate>Mon, 17 Dec 2018 14:15:29 GMT</pubDate>
            <atom:updated>2018-12-17T14:15:27.925Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UL7RccGu6xXYCsBTcKtTWg.jpeg" /></figure><p>There was a time (some still do) when companies that built tools (Apple, Microsoft) used to target business owners, software engineers, students, creatives and artists via their advertising because those particular industries were taking off.</p><p>And now, companies have started to target influencers directly with their advertising. Below is a recent ad from the Sunday edition of Times of India where HP is trying to appeal to influencers &amp; creators specifically.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tLX3OOwJiXwdSNseZ0qWeA.jpeg" /></figure><p>This may seem like something small but it says a lot about an industry that’s taking off</p><p>Going forward, influencer marketing and hence influencers are going to be a key part of a brand’s marketing strategy and they cannot be ignored.</p><h4><a href="/service/mailto:hello@winkl.co">Need help with influencer marketing? Talk to us.</a></h4><blockquote>Found this useful? Please clap and subscribe below. This ensures more people like you will be able to derive value from this article!</blockquote><p><strong><em>At Winkl, we’re building a platform for content creators that ensures creating content is rewarding and isn’t lonely. Winkl is used by more than 3500 bloggers &amp; influencers to share their work, get feedback, inspiration, community and brand collaborations.</em></strong></p><p><a href="http://winkl.co/">Join Bloggers, Influencers and Creators from across the world.</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ad83e61bebe0" width="1" height="1" alt=""><hr><p><a href="https://medium.com/winkl-insights/the-influencer-wave-is-coming-ad83e61bebe0">The influencer wave is coming!</a> was originally published in <a href="https://medium.com/winkl-insights">Winkl Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Move over celebs! Micro-Influencers are here (>15% CTR!)]]></title>
            <link>https://medium.com/winkl-insights/move-over-celebs-micro-influencers-are-here-15-ctr-e9b54ae4b90d?source=rss----f9a703a784dc---4</link>
            <guid isPermaLink="false">https://medium.com/p/e9b54ae4b90d</guid>
            <category><![CDATA[wired]]></category>
            <category><![CDATA[product]]></category>
            <category><![CDATA[micro-influencers]]></category>
            <category><![CDATA[influencer-marketing]]></category>
            <category><![CDATA[winkl]]></category>
            <dc:creator><![CDATA[Thakur Rahul Singh]]></dc:creator>
            <pubDate>Mon, 10 Dec 2018 14:56:29 GMT</pubDate>
            <atom:updated>2018-12-10T14:57:51.204Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pKb6CiJm2puBu8sUqZaUhw.jpeg" /></figure><p>Yes! It’s the age of the micro-influencers and they are going to be the future of influencer marketing. In 2018 and going forward, brands are preferring to work with many micro influencers instead of a big name for their campaigns. And this behaviour is only going to strengthen.</p><h3><strong>So who are micro influencers…</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*6HR3qogJj2doeedIgvUucQ.jpeg" /></figure><p>Micro influencers are people who have a social media reach of about 10,000 to 100,000 people. They are mostly regular people like you and me but by virtue of them publishing good content, they gain an audience base over time.</p><blockquote>But because they do not have a celebrity status as such and are actually content creators &amp; story tellers more than anything else, their audience tends to be highly engaged and loyal.</blockquote><p>This key aspect is being leveraged by brands and agencies to drive real ROI instead of simply hiring a “well known face” as the brand ambassador/influencer.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/900/1*EiMoq1VKGwx6M4EbJ7MwDg.png" /></figure><p>Micro influencers enjoy a huge amount of trust in the smaller networks they are present in. This may not necessarily drive huge reach, but the conversions are very high.</p><h3>CTR of &gt;15%!</h3><p>In-fact, in one of our recent campaigns, micro influencers were able to generate a CTR of more than 15%! <strong>That’s more than 30x the industry average!</strong> <em>(a separate blog post on this later)</em></p><p>When we started Winkl about a year and half ago, the below was one of our major hypotheses,</p><blockquote><strong>In the future, brands will work with many micro influencer instead of 1 celeb or a few macro influencers as they will drive higher engagement and ROI.</strong></blockquote><p>Very recently, a leading authoritative publication on consumer internet, Wired,<strong> published a story that supported our hypotheses </strong>about the shifting behaviours and trends in the influencer marketing space!</p><p><strong>Checkout the article below —</strong></p><p><a href="https://www.wired.com/story/logan-paul-micro-influencers-youtube-instagram/">Byeeeee, Logan Paul: Brands Prefer &#39;Micro Influencers&#39; Now</a></p><p>We feel that we’re in a space that’s undergoing fast changes and disruption. This means large opportunity ($100bn) and we’re super excited to play our part in shaping the future of influencer marketing in India and around the world.</p><p>When there are many micro influencers to potentially work with, how do you pick the most relevant ones? How do ensure you communicate to them effectively at scale? How do you track performance and ROI from so many influencers in real time? This is how Winkl helps marketeers with.</p><p><a href="https://calendly.com/winkl/15min">Need help with this? Talk to us!</a></p><blockquote>Found this useful? Please clap and subscribe below. This ensures more people like you will be able to derive value from this article!</blockquote><p><strong><em>At Winkl, we’re building a platform for content creators that ensures creating content is rewarding and isn’t lonely. Winkl is used by more than 3500 bloggers &amp; influencers to share their work, get feedback, inspiration, community and brand collaborations.</em></strong></p><p></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e9b54ae4b90d" width="1" height="1" alt=""><hr><p><a href="https://medium.com/winkl-insights/move-over-celebs-micro-influencers-are-here-15-ctr-e9b54ae4b90d">Move over celebs! Micro-Influencers are here (&gt;15% CTR!)</a> was originally published in <a href="https://medium.com/winkl-insights">Winkl Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why we decided to use Ruby on Rails for Winkl]]></title>
            <link>https://medium.com/winkl-insights/why-we-decided-to-use-ruby-on-rails-for-winkl-712145f5c364?source=rss----f9a703a784dc---4</link>
            <guid isPermaLink="false">https://medium.com/p/712145f5c364</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[framework]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[ruby-on-rails]]></category>
            <dc:creator><![CDATA[Nikhil Kumar]]></dc:creator>
            <pubDate>Mon, 10 Dec 2018 13:01:45 GMT</pubDate>
            <atom:updated>2018-12-10T13:01:45.781Z</atom:updated>
            <content:encoded><![CDATA[<h4>What we looked at before deciding to use Rails as the primary framework.</h4><h3>About Winkl</h3><p>Winkl is a platform that helps influencers create their profile, showcase their content, get feedback from other influencers and make money via brand collaborations.</p><p>It ensures that content creation for influencers is emotionally &amp; financially rewarding journey and not a lonely one.</p><p>Winkl on the other side helps brands run influencer marketing campaigns in a way that’s as easy to run a Facebook ad.</p><h3>Introduction</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*P6CO-kt_dt7wc9m-A-c7Yg.jpeg" /></figure><p>As a startup, it is most important for you to get things out ASAP, get feedback and iterate. The speed at which your product gets to the hands of your customers is key. This is where Ruby on Rails helps.</p><p><a href="https://rubyonrails.org/">Rails</a> is a development tool which gives developers a framework for all the code they write. The Rails framework helps developers to build websites and applications, because it abstracts and simplifies common repetitive tasks.</p><p><strong>Ruby on Rails is open source software</strong>, so not only is it free to use, you can also help make it better. <a href="http://contributors.rubyonrails.org/">More than 4,500 people</a> already have contributed code to Rails since 2004.</p><h3>Backstory</h3><p>Prior to picking and building with RoR(Ruby on Rails) I developed backend systems using PHP and Django. It seemed mediocre and did not work well for us.</p><p>Before writing the first line of code, I started off by picking out the top 3 frameworks out there, learning them and building a small app using each one of them just to get a taste of it. I picked <strong>Express.js, Ruby on Rails and Django</strong>(with some existing knowledge) for the task. I started off with each of the frameworks by building out the small app. At the end of it I was blown away by the speed and structure offered by RoR.</p><p>I was able to finish the app in <strong>half the time</strong> compared to the other frameworks. The deployment was not only quick but the code was extremely well structured. A new person joining the team will be able to pickup, learn and contribute immediately.</p><h3>Advantages of RoR:</h3><p><strong>Super Fast Development</strong> — You can get started with a brand new Rails project with just one line of code. You get the whole MVC framework setup for you right out of the box. Development and pushing code is extremely quick which is a great choice for startups looking to push code ASAP.</p><p><strong>East to Learn</strong> — It is very simple to learn. With around 2–3 hrs of courses you should be able to get to work on your first app.</p><p><strong>East to Pickup</strong> — Since the framework structure is built so well it is easy for new employees to learn, understand and contribute to the code quickly.</p><p><strong>Gems</strong> — The community probably has a gem for most of the issues you need to solve. You can get the gem into your project with just a couple of lines of code.</p><p><strong>The community</strong> — The community is really strong and willing to help. Lot of open source material and courses available to help you. <a href="https://gorails.com/">https://gorails.com/</a>.</p><h3>Disadvantages of RoR:</h3><p><strong>Resource Consumption</strong> — RoR needs a lot of CPU usage and RAM</p><p><strong>Large Projects</strong> — Large projects need to be coded well to make sure it works fast. After some complexity, performance on RoR may not be as great as some other frameworks.</p><p><strong>Not all hosts support RoR deployment — </strong>However you can still use hosts like AWS, Heroku and EngineYard to host your RoR code.</p><h3>Our Decision</h3><p>After considering all the points above and at the time, we were building an app which allowed people to have their own subdomain/domain which used multiple routing mechanisms. As this was one of the core requirements for the app, Rails had exceptional support for routing and it was very well structured.</p><p>We made the decision to use RoR to get things going. To our surprise we were able to build out our MVP in 3 weeks and get a working product out to our customers. We got immense feedback and kept iterating. We had pushes going out every day. The speed of development was just amazing. In just a month after launching we had our first customer subscribing for a 2 year plan. It felt like we are on to something.</p><p>Currently after about a year I am starting to see that it does consume a lot of RAM and CPU power. But it is possible to tune the app and the development process to minimise these problems (More on that in a future blog).</p><p>Many large companies have been built using RoR and without it they would not have been able to make it.</p><p>Some Companies built on RoR:</p><ol><li>Basecamp</li><li>Airbnb</li><li>Github</li><li>Twitter</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LVjMtuQmbmtzN-qhzULy0w.png" /></figure><p>Rails has been an amazing journey and it has helped businesses of all sizes.</p><p>You can keep up with Rails <a href="https://twitter.com/rails">here</a>(Twitter) and <a href="https://weblog.rubyonrails.org/">here</a>(blog).</p><blockquote>Found this useful? Please clap and subscribe below. This ensures more people like you will be able to derive value from this article!</blockquote><p><strong><em>At Winkl, we’re building a platform for content creators that ensures creating content is rewarding and isn’t lonely. Winkl is used by more than 3500 bloggers &amp; influencers to share their work, get feedback, inspiration, community and brand collaborations.</em></strong></p><p></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=712145f5c364" width="1" height="1" alt=""><hr><p><a href="https://medium.com/winkl-insights/why-we-decided-to-use-ruby-on-rails-for-winkl-712145f5c364">Why we decided to use Ruby on Rails for Winkl</a> was originally published in <a href="https://medium.com/winkl-insights">Winkl Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Milestone #1 — The first customer.]]></title>
            <link>https://medium.com/winkl-insights/milestone-1-the-first-customer-23c4d3234113?source=rss----f9a703a784dc---4</link>
            <guid isPermaLink="false">https://medium.com/p/23c4d3234113</guid>
            <category><![CDATA[growth]]></category>
            <category><![CDATA[first-customers]]></category>
            <category><![CDATA[winkl]]></category>
            <category><![CDATA[milestones]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[Thakur Rahul Singh]]></dc:creator>
            <pubDate>Fri, 07 Dec 2018 12:29:49 GMT</pubDate>
            <atom:updated>2017-06-18T16:37:23.875Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WmFwXmI_Km3Kwr8EVCIxFA.png" /><figcaption>The first customer is really hard. And very important too.</figcaption></figure><p>Building a startup into a real business is hard. And often, an important milestone is getting your first customer. That is when the rubber meets the road, and things get real.</p><p>And it’s even harder when you’re building something new and different. For example in most service based businesses like web development or restaurants, getting customers isn’t that hard because people are used to paying for those services. But when you’re building a new product that’s different from what exists, getting that customer is difficult. It’s so difficult that tons of resources like <a href="https://yourfirstcustomers.com/">these</a> exist to help you get there.</p><blockquote>We got our first customer about 2 weeks ago and the feeling was surreal. The feeling that someone was actually paying us for what we built, in what we put long lonely hours into. The moment the first money hit the bank, I knew we were onto something. Something that people will actually pay us for.</blockquote><p>It took us about 10 months to get here. With a pivot and many mistakes along the way. We started off fresh after college with no experience at all. At times it felt scary and stupid. But I guess it’s all about the milestones, that’s fulfilling more than anything else.</p><p>What helped us get the first customer was a blogger meet-up we organised called <a href="http://winkl.co/meets"><strong>Winkl Meets.</strong></a><strong> </strong>Where we pitched our product to users in person. It was manual and old school. Nothing fancy! But that’s all it takes to validate what you’re building.</p><p><strong>So one learning that I got out of this is —</strong></p><blockquote>Before you write that first line of code and build your product, make sure there’s someone who’s going to pay you for it.</blockquote><p>It couldn’t get more exciting for me that it is currently as we go about building <a href="http://winkl.co/">Winkl</a>. It’s an adventure that I’ll have the most amount of fun on.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=23c4d3234113" width="1" height="1" alt=""><hr><p><a href="https://medium.com/winkl-insights/milestone-1-the-first-customer-23c4d3234113">Milestone #1 — The first customer.</a> was originally published in <a href="https://medium.com/winkl-insights">Winkl Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[1000 Creators on Winkl | A reach of 3 million people — One year!]]></title>
            <link>https://medium.com/winkl-insights/1000-creators-on-winkl-a-reach-of-3-million-people-one-year-de3740dd2639?source=rss----f9a703a784dc---4</link>
            <guid isPermaLink="false">https://medium.com/p/de3740dd2639</guid>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[milestones]]></category>
            <category><![CDATA[winkl]]></category>
            <category><![CDATA[growth]]></category>
            <dc:creator><![CDATA[Thakur Rahul Singh]]></dc:creator>
            <pubDate>Fri, 07 Dec 2018 12:26:31 GMT</pubDate>
            <atom:updated>2018-12-07T12:28:20.764Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ufGkremhWW1HEyhV8a8aaQ.jpeg" /><figcaption><a href="http://winkl.co/1000-milestones">A creative from a recent campaign that we ran</a></figcaption></figure><p>A little over a year ago, on 7th May 2017, we released the first version of <a href="http://winkl.co/">Winkl</a> (in its current form). Prior to that for about 12 months, <a href="https://medium.com/u/1dacced4cad2">Nikhil Kumar</a> and I tried a bunch of things (<a href="https://www.thehindu.com/news/national/karnataka/four-students-of-nitk-surathkal-bring-out-winkl-the-shopping-app/article8028839.ece">a proximity based mall shopping app</a> and fashion Q&amp;A platform) with hardly any success. <strong>When we shut down both of those, no one really cared.</strong></p><p>In March 2017, we had about 4 digits of money (INR!) left in the bank and were days away from shutting down. But for some reason, maybe due to stupidity or unrealistic levels of optimism, we decided to give it one last shot. <strong>One last try to build something we felt we had a good insight on and something that we personally cared about.</strong></p><p>In the previous months of building products and trying to get traction, we ended up interacting with many content creators (bloggers, writers, influencers etc.) who’d always keep telling us about the issues they faced. They couldn’t setup their website/blog easily, they couldn’t get traffic, couldn’t monetise etc., and this list for long. <strong>For us, the solution was obvious and we wondered why hadn’t anyone done this before</strong>. And the more we spoke to creators, it was crystal clear that something needed to be done.</p><p>To keep the ship afloat and have some money coming in, we started doing consulting projects, creating websites, &amp; presentation decks and what not. While we did these projects, we started building Winkl on the side.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Q4OwfP4FTEazlU_So7ZXwQ.jpeg" /><figcaption>The first ever mockup of a Winkl blog</figcaption></figure><p>It was hard. A lot of our time was spent in meeting clients for these projects, doing what they wanted us to, listening to their qualms etc. We hardly got any time to work on Winkl. But we kept at it.</p><p>Within a month from the first mockup, on <strong>7th May 2017</strong>, we had something ready. It was a basic product where anyone could create their blog that looks good in under a minute.</p><p>We showed it to a few people — they really liked it and started using it. And the only question they ever asked — <em>What if you shut down?</em></p><p><strong>We didn’t have an answer, but we were in business.</strong></p><p>We knew that if we shut this down, there’d be people we were answerable to. Within a month, on 8th June 2017 (coincidentally my birthday) we had our first paying customer.</p><p>The feeling was unreal. It gave us conviction to keep going. It validated to an extent that our show of stupidity wasn’t so stupid after all.</p><p>Since that day, we’ve come a long way. We’ve raised some angel funding from some amazing investors and advisors who’re more of friends and partners. We’ve received great feedback from our users, worked with <a href="http://winkl.co/ylg-squad">top brands</a> in India, we’ve been part of the <a href="https://inc42.com/buzz/meet-the-19-startups-selected-for-the-summer-2018-accelerator-cohort-of-axilor/">amazing batch at Axilor</a>, got to work with <a href="https://angel.co/itsonwinkl/jobs">some superb people</a>, made some mistakes, learnt a lot and so much more.</p><blockquote><strong>A couple of days back, we hit the mark of 1000 creators on Winkl who have a cumulative reach of more than 3 million people! It then dawned upon us about the amazing journey we’ve been privileged to be a part of.</strong></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/452/1*Zbm7jDdXLP4TtdlJ682MaA.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/427/1*sbbSNTMCVVQxzQjbr-qFNw.jpeg" /><figcaption><strong>Some of the feedback from our users!</strong></figcaption></figure><p>While we continue to realise our mission of helping creators across the world succeed in whatever they choose to create, <strong>I couldn’t help but wonder, what if we had given up? What if we never gave that one last shot?</strong> Thinking about this makes me feel good about being stupid and unreasonable.</p><p>The more we work towards our vision, the more there is to do. Which means, we’re almost always 1% done and we’re just getting started. The fact that we’ll possibly never be done, is a great motivator for us.</p><p>Thanks to our users, brand partners, mentors, investors, friends &amp; critics and out team, for being with us. If it wasn’t for you, we wouldn’t have made it this far. With you around, the journey so far has give us more than what we could ever ask for.</p><p>Here’s to many, many more milestones and years ahead.</p><blockquote>Found this useful? Please clap and subscribe below. This ensures more people like you will be able to derive value from this article!</blockquote><p><strong><em>At Winkl, we’re building a platform for content creators that ensures creating content is rewarding and isn’t lonely. Winkl is used by more than 3500 bloggers &amp; influencers to share their work, get feedback, inspiration, community and brand collaborations.</em></strong></p><p></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=de3740dd2639" width="1" height="1" alt=""><hr><p><a href="https://medium.com/winkl-insights/1000-creators-on-winkl-a-reach-of-3-million-people-one-year-de3740dd2639">1000 Creators on Winkl | A reach of 3 million people — One year!</a> was originally published in <a href="https://medium.com/winkl-insights">Winkl Insights</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>