<?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 Pritha Saha on Medium]]></title>
        <description><![CDATA[Stories by Pritha Saha on Medium]]></description>
        <link>https://medium.com/@prithasaha_26820?source=rss-6b926c2e956e------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*FsheT6zQcBryeui3hWsJxw.jpeg</url>
            <title>Stories by Pritha Saha on Medium</title>
            <link>https://medium.com/@prithasaha_26820?source=rss-6b926c2e956e------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 12:11:06 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@prithasaha_26820/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[Why we sell the way we do?]]></title>
            <link>https://medium.com/fynd-team/why-we-sell-the-way-we-do-89a607388885?source=rss-6b926c2e956e------2</link>
            <guid isPermaLink="false">https://medium.com/p/89a607388885</guid>
            <category><![CDATA[ecommerce]]></category>
            <category><![CDATA[omnichannel]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[growth]]></category>
            <dc:creator><![CDATA[Pritha Saha]]></dc:creator>
            <pubDate>Tue, 12 Jun 2018 04:31:01 GMT</pubDate>
            <atom:updated>2018-06-12T05:28:31.520Z</atom:updated>
            <content:encoded><![CDATA[<p>A deeper look at what we could sell versus what we already sell</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*K1CFkY-2PK5VGsWNqdUOuw.jpeg" /></figure><h3>Beating the Pareto psychology</h3><p>Vilfredo Pareto, an economist, had made an observation back in the 1900s, that 20% of the Italian population owned 80% of the property. This groundbreaking observation was later adapted to most business problems and is a popular tool to identify the 20% input that is causing the 80% output.</p><p>Pareto’s Principle implies that most of our effort should be directed towards those 20–30% inputs which generate the most output. However a lot of businesses have proven this wrong. Netflix gets most of its profits from a large number of infrequently requested movies rather than a few blockbuster and profitable ones. The same holds true for songs bought on Itunes or products bought on Amazon. This is where the theory of <strong><em>Long Tail</em></strong> comes in. While Pareto’s theory can be applied only where there is a known workflow and applies in hindsight, Long Tail tells us how we can reap profits by selling “More of less”.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DenPYwdnbppTnNXl1jYtEA.jpeg" /></figure><p>Although a small percentage of the products contribute to the most popularity, the cumulative sum of popularity of the larger percentage of products cannot be discarded either.</p><p>According to Chris Anderson, author of the book <strong><em>The Long Tail: Why the Future of Business Is Selling Less of More</em></strong>, our “economy and culture is shifting from mass markets to millions of niches”.</p><p>Traditional retail economics dictate that only hits get stocked in the store. But in an online world, where we pay zero cost to showcase inventory, there is ample scope to stock virtually everything. It is ultimately the millions of niche choices, which cater to people’s narrow interests.</p><p>We found this theory to be true for our business as well, as demonstrated by the below graph:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*goQYf1lTTd721cvW_7QDyw.jpeg" /></figure><p>This graph shows the sales contribution of all brands within a 3 month period.</p><p>We found that minimal additional efforts are going behind the small percentage of brands which are generating most of the sales, just like hit movies on Netflix.</p><p>As a platform, that strives to provide the best of fashion to consumers, we decided to focus on the width of our assortment which are live on our platform and have the potential to generate much more cumulative revenue, in other words the long tail of our assortment.</p><p>We discovered a variety of issues with the long tail, ranging from commercial issues, low percentage of fresh content, pricing issues, frequent order cancellations, product defects, inventory sync issues, delays in order fulfillment, etc.</p><p>Since these problems were interlinked, it was difficult to decide which one to prioritise first. After several rounds of brainstorming, we came up with a process of scoring brands on the basis of different parameters, which could serve as a reference to the category managers to come up with a priority order. This was how we got started with our Brand Audit Methodology.</p><h3>Structuring the Project</h3><p>The Project was broken down into three parts:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/708/1*yTDk6bakBbr_EpD-h2MGDw.png" /></figure><h4><strong>1. Gathering Data</strong></h4><p>The first task was to get all the brand metrics in a single place. This proved to be an extensive task as data was not maintained in the format that was needed. For instance, all metrics were categorised brand wise, however there were many instances of one brand coming from multiple companies. One would be the parent company and the others would be franchisees or distributors. In that case, for each company-brand combination, the performance and the commercial terms would be different.</p><p>Once the data structure was resolved, we got down to defining the metrics. The tool we used in this regard was <strong><em>Google BigQuery</em></strong>, to link data from several database and transform data in the way we wanted to view.</p><p>The data points were grouped under the following heads-</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/654/1*55euKIzqpsgj6bwWHWlY1g.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/636/1*nYIe1ARxQFdMX1twuaGJXA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/647/1*E94PVciaIwSzRZtWEdBsuw.png" /></figure><p>All in all there were around 26 metrics for a single company-brand.</p><p>In order to remove any legacy bias, we were always tracking these metrics on a rolling 90 days period. The metrics were always viewed in percentage terms and not absolute figures.</p><h4><strong>2. Analysis</strong></h4><p>Now we had a mammoth set of data, with everything in place. But it was still difficult to make out what was happening with a particular company-brand. For instance sales might be high for the brand, but the product returns are huge ultimately generating negative ROI. Also, we didn’t know how to compare performances. Brand A could be priced at Rs.500 and Brand B could be priced at Rs.1,000, but with a 50% discount. If Brand B generated more sales than Brand A, would it be easy to conclude that this is because the perceived value of Brand B is more? Or is that just one of the factors?</p><p>We proceeded to study the factors , which could influence purchases, as outlined below</p><p><strong>Case 1: </strong>In case a brand was coming from different retailers, the pricing and the nature of products could deviate across the companies. For instance, in case of a particular lifestyle brand, we were live through two companies — parent company and a distributor. Our selling pattern was different for both of them as per the below snapshot of all the metrics’ data:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/665/1*n6xXiV5BsPLHJ2NO82jEZg.png" /></figure><p><strong>Case 2:</strong> Gender preferences were varied when it came to pricing. Women were more price sensitive than men, hence the average selling price in certain categories were more for men than women. For a particular footwear brand this was the average selling price for a particular month:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rJ_I_y2juIhrVu6Bm_bxQQ.jpeg" /></figure><p><strong>Case 3</strong>: Behaviour varied for every category. Making a decision to buy a Rs.500 top was far easier than buying a Rs.500 necklace. There were three levels of categories — L1, L2, L3. L1 would be the broadest category, like Clothing. L2 would be a sub category like Dresses and Jumpsuits and L3 would be a sub category within L2, like Skater Dress. Price sensitivity was mostly within the L2 category, as demonstrated below. Hence we decided to consider L2 as a factor.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/669/1*XBnSx2F_kS-sth6UZ2wcmw.png" /></figure><p>Basis our studies, we decided to divide our entire assortment into segments, which are nothing but <strong><em>“company-brand-gender-category”</em></strong> combinations. There are almost 2,000 such combinations in Fynd.</p><p>Now we needed to find and segregate similarly behaving segments to make a fair comparative assessment between them. We decided to place similar segments in a price bucket. This price needed to be the representative price of the collection of buckets</p><p><strong>Segregating similar segments/datasets (based on price)</strong></p><p>This was divided into two parts:</p><p>a) Identifying each segment/dataset on the basis of a parameter (price in this case). Firstly we calculated the average price across all items in a segment. However, the average price did not seem to be a good indicator for the segment, as in certain cases the deviation amongst prices could be high. So we calculated the median price of the segment. Lastly, we calculated the minimum of the average and median price. This seemed to be a fair indicator of the segment price. We will be referring to this representative price as min_price from now on.</p><p>Below table showcases the min_price for different segments, for a particular brand:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/662/1*8HA7pouaF3LXrXuglnVHvQ.png" /></figure><p>b) Bucketing similar segments/datasets</p><p>This was a slightly subjective decision and was decided after studying customer profiles and a round of brainstorming. For segments, whose min_price was below Rs.2,000, the price sensitivity seemed to change for every Rs.400. For segments whose min_price was below Rs.4,000, the price sensitivity seemed to change for every Rs.1,000. Between Rs.4,000 and Rs.8,000, the price sensitivity seemed to change for every Rs.2,000.</p><p>When we say the price sensitivity changes within a particular price segment, it means that items within a certain range will probably be evaluated equally. Hence a customer will view a t-shirt priced at Rs.200 and a t-shirt at Rs.300 on the same plane. Similarly a customer will view a dress valued at Rs.5,000 and at Rs.7,000 on the same plane.</p><p>We formulated the below logic for assigning the price buckets, or Price_Round, as we shall be referring from now on:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/1*L_IO1Ut0b9p8hIuZSotH0g.png" /></figure><p>We decided to review this logic every quarter.</p><p>Based on this logic, we now had the respective Price_Round for each segment.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/703/1*HuEdrIy1yat77ocnfnjWVA.png" /></figure><p>So within a particular Price_Round, in a certain category and for a particular gender, we could compare the performances of each company-brand and evaluate them.</p><h4><strong>3. Representing Data</strong></h4><p>Now that we had put a method to the madness, the next step was to represent the data in a way that was easily and quickly comprehensible. There were a lot of metrics we were tracking, hence making it difficult to immediately conclude if the health of the brand was good or bad.</p><p>We decided to select some crucial metrics, put certain weightage and formulate a score. These metrics were again normalised to depict the comparative performance within a price band.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/658/1*ECy8RlJ4YkvzlwdVQ6uCAQ.png" /></figure><p>Below scoring chart was used:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/850/1*EA9E1c1XWWrrVaZqAwdkcg.png" /></figure><p>The weightage was assigned between absolute values of 1 to 2, to give similar importance to all metrics. Factors influencing sales/business the most have been assigned a weightage of 2, followed by 1.5 and 1. Assigning higher numbers to a very critical metric runs the risk of biasing the score towards a particular attribute, while undermining others. However this would be under observation for the first three months of the project and then could be subject to change based on the scenario.</p><p>The metrics marked in blue were calculated together to form <strong>Orders Score</strong>.</p><p>The metrics in green were calculated together to form <strong>Content Score</strong>.</p><p>The metrics in pink were calculated together to form <strong>Traffic Score</strong>.</p><p>The threshold score for each was the average score in that segment.</p><p>Based on this calculation, we now had a clear picture of performance among segments:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*L0a9JzL6QNgvNLrJOuuh1A.png" /></figure><p>This is the score card for one of the men’s sub categories within Footwear. The brand names have been hidden for confidentiality purposes. For each company-brand combination, we now know the performance in each sector-Orders, Content and Traffic, with respect to a particular Price_Round.</p><p>For instance within the Rs.2,000 Price_Round, one company-brand is performing above threshold in Orders, but below threshold in Content and Traffic. There is one brand, which is performing below threshold in all 3 sectors. This is a warning signal and pretty much indicates that it needs the category manager’s attention.</p><p>Now to identify the exact nature of the problem, the category manager can deep dive into the individual metrics of the particular sub category:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/936/0*793zPP31Cjw_Ujvb." /></figure><p>This is just a snapshot of the entire metrics board. All the metrics here were only applicable to segments.</p><p>For the entire visualisation, we used Bime Studio, which we connected with data from Google BigQuery.</p><p>With this, we finally concluded laying the foundation for our Brand Audit project.</p><p>However, this is far from over. We would need the support of our brand partners to achieve our joint goals together. This cannot happen overnight by enforcing a new process.</p><p>Hence we have thought of a creative approach to engage our brand partners through a human focussed design, or in other words a gamified journey, to attain goals and earn rewards on the way. This is still getting conceptualised.</p><p>For now we need to keep our sights focussed on the metrics that define our vision of the <strong><em>Long Tail</em></strong>.</p><h4><strong>References</strong></h4><ul><li><a href="http://www.longtail.com/about.html">The Longtail</a></li><li><a href="http://epilepsygenetics.net/2013/03/14/the-pareto-principle-versus-the-long-tail/">The Pareto Principle versus the Long Tail</a></li><li><a href="https://betterexplained.com/articles/understanding-the-pareto-principle-the-8020-rule/">Understanding the Pareto Rule</a></li></ul><p><em>We welcome your feedbacks, insights and suggestions for topics that you would like us to cover, please reach out to us at </em><a href="/service/mailto: brands@gofynd.com"><em>brands@gofynd.com</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=89a607388885" width="1" height="1" alt=""><hr><p><a href="https://medium.com/fynd-team/why-we-sell-the-way-we-do-89a607388885">Why we sell the way we do?</a> was originally published in <a href="https://medium.com/fynd-team">Building Fynd</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 buy what we buy?]]></title>
            <link>https://medium.com/fynd-team/why-we-buy-what-we-buy-505bab2ede0?source=rss-6b926c2e956e------2</link>
            <guid isPermaLink="false">https://medium.com/p/505bab2ede0</guid>
            <category><![CDATA[marketing]]></category>
            <category><![CDATA[online-sales-funnel]]></category>
            <category><![CDATA[customer-engagement]]></category>
            <category><![CDATA[ecommerce]]></category>
            <dc:creator><![CDATA[Pritha Saha]]></dc:creator>
            <pubDate>Thu, 25 May 2017 06:05:29 GMT</pubDate>
            <atom:updated>2017-05-25T06:05:29.128Z</atom:updated>
            <content:encoded><![CDATA[<p>As a fashion marketplace, we at <a href="http://www.gofynd.com/">Fynd</a> are constantly striving to showcase better brands and better assortment to consumers. Purchase decisions, though can be a tricky process. After hours of evaluation, only about 5% of them might end up in the shopping carts, and of them around 15% actually get placed. This would still be the best case scenario depending on the brand recall, price point, product information depth, etc.</p><p>The traditional marketing funnel looked something like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/671/0*EAbwc074bMYbhDWQ." /><figcaption>Source: <a href="http://www.mckinsey.com/business-functions/marketing-and-sales/our-insights/the-consumer-decision-journey">The consumer decision journey (Mckinsey,2009)</a></figcaption></figure><p>This has now been rightly ditched by marketing experts, as it does not take into account the two way conversation between the brand and the consumer. Imagine in case of Fynd, people first getting aware about our proposition, but at the same time going through online reviews and comparing prices while browsing through the app! These complex interactions have now deemed this funnel dead! Instead an article from McKinsey suggests the below model as a purchase journey:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/695/0*G288AcUZqDFn1lGC." /><figcaption>Source: <a href="http://www.mckinsey.com/business-functions/marketing-and-sales/our-insights/the-consumer-decision-journey">The consumer decision journey (Mckinsey,2009)</a></figcaption></figure><p>While the first stage is taken care of by the customer acquisition team, once we have the audience in the app, we can play a pivotal role in the “active evaluation process”, contrary to what a consumer might believe to be her own empowered purchase.</p><p>Analogous to real estate property, there are certain prime areas of the app or website which attract more eye balls than the rest. Hence product placement in those particular areas inevitably yield an influencing power on the consumer psyche. The most visible products act as a bait to attract product views to that particular brand and category, leading to add-to-carts and ultimately purchases.</p><p>Playing with visibility gives us a way to play with individual brand funnels and influence metrics like conversion, awareness etc. In our case the prime areas are the “Feed”, “Brand Sort” and the “Collections” sections. Consumers are drawn to these particular areas due to the specific construct of the app, hence most purchases happen from here.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/616/1*hSTeCg4wdItgNy67zMw_sA.png" /><figcaption>Multiple properties on the Fynd app that we utilize to actively engage consumers</figcaption></figure><p>Buildings in prime real estate areas are valuable because they provide ease of connectivity to airports, malls and schools among other thins. Also, they are always better than buildings in non prime areas in terms of construction and layout, in order to uphold the image of the area. Likewise, brands placed in the prime areas have an onus of leading to the brand page and converting to sales and they also project the stature of the marketplace to a fashion conscious consumer. We have observed conversion ratios almost doubling for brands who have got a traffic boost. Placement in the prime areas also helps in discoverability of brands leading to purchases even after removing them from the prime areas after a few days of consistently boosting them.</p><p>Leveraging this theory, we prioritise certain brands on a weekly basis, which fulfil our business objectives as well as help us deliver a good customer decision making and shopping experience.</p><p>We designed the below evaluation parameters to aid us in selecting these particular brands:</p><ol><li>Study of men’s and women’s categories with respect to number of views received last week, followed by selection of two low performing categories and one high performing category based on the same. This was essentially done to ensure that the low performing categories get a visibility boost and the high performing category helps us in the overall traction and maintains the volume.</li><li>Selection of two brands in each category based on the respective criteria:</li></ol><p>a) Decent conversion from “views” to “add to carts”, ranking midway in terms of total views. The hypothesis is that this brand would probably sell more if we give it enough traffic, since it has an inherent decent conversion.</p><p>b) Low ranking in terms of traffic received as well as less than average conversion. This could be a function of many factors like pricing, content, etc. The audience might have lost interest in browsing this brand on our app. Hence even after updating a new collection of the brand we might not generate sales. In this case the hypothesis is that we need to rekindle interest in the brand to improve the brand health.</p><p>The below experiment on brand sales was carried out for a menswear brand. We chose the category “Men’s Shirts” as one of the top ranking categories with respect to traffic and chose a brand based on the first criteria-medium views with a decent conversion of 4% from views to add to carts.</p><p>Initially the brand was under an observation period from Day 1 to Day 7. The “view share” denotes the share of overall traffic given to the brand, “Conversion” denotes add to carts from views, “Brand Contribution” denotes the business contribution of the brand to the category.</p><p>The infographic shows the effect on brand metrics when we increased the view share, starting from Day 8. “Add to cart” conversions from views increased. But more importantly, these “add to carts” got converted to actual sales and helped boost up the brand contribution. Before Day 8 consumers were adding to their carts but not really placing the final order. It is the increased visibility of the brand which finally helped them to make their purchase decision. The point of inflection clearly happened on Day 8, when the yellow line crossed the red line. Also it is interesting to note that the brand contribution dropped proportionately when the view share dropped on Day 12 and then stabilised again on increasing the same.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/872/0*56YLC8O4yA7NtyZ0." /><figcaption>Change in Brand Metrics with increase in traffic</figcaption></figure><p>In this case views are mostly attributed to the Feed page as shown in below graph.</p><p>On Day 8, which is 31st Mar, we gave a slight boost to the traffic through Brand Sort. Thereafter it was predominantly feed that attributed to the huge and proportionate increase in the view share. Hence the feed gives us control of the driver’s seat from where we can influence the purchase choices of the consumers.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*wvLZVvn6UYfimsbmsIdcGw.png" /><figcaption>Traffic Volume by Source</figcaption></figure><p>As a result of this experiment we were also able to boost the overall category by 69%. This is a crucial finding as it denotes that some brands can be appointed as category leaders and help customers in their evaluation phases.</p><p>In another instance we observed the effects on a brand, post removing it from a prime property. The brand in question belonged to a low performing category-”Women’s accessories”. The brand was low ranking in terms of overall view share and had below average conversion.</p><p>However we had just uploaded the SS17 collection of the brand and decided to use the new assortment as a means of arousing consumer interest. Here again we gave the initial boost through “Feed” to showcase the assortment and observed the results post the period.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/783/0*qTF67yvmR3WQeVgK." /><figcaption>Traffic Volume by Source</figcaption></figure><p>We started giving a visibility boost from 1st May till 6th May, then gradually decreased the views. However the audience got redirected to brand and category search pages for the products, and we ended up increasing the conversion percentage even with less view share.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/847/0*tZBBqeir0YpQIIu6." /><figcaption>View Share vs Conversion</figcaption></figure><p>This is the kind of brand which could essentially be used to sustain loyalty of consumers, who now believe that we always have the freshest fashion.</p><p>To conclude this experiment helps us to identify the behaviour of brands and tweak our efforts accordingly to influence different areas of the funnel.</p><p>However how does one scale up and make decisions when there are countless brands but only limited space to drive visibility?</p><p>This is where we can take advantage of the “high demand and low supply situation” and start monetizing our spaces. This is very similar to how property prices escalate when demand rises in a particular area and no one has any qualms in investing money as they are sure of the results — in terms of sales or eyeballs.</p><p>In fact this would be the perfect win-win situation for us, as we influence different touch points of the purchase funnel and for the brand as they get the assured visibility leading to sales.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=505bab2ede0" width="1" height="1" alt=""><hr><p><a href="https://medium.com/fynd-team/why-we-buy-what-we-buy-505bab2ede0">Why we buy what we buy?</a> was originally published in <a href="https://medium.com/fynd-team">Building Fynd</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>