3 Blunders We Made While Building Fynd Over The Last 1 Year

Fynd completed one year as a consumer internet company on October 20, 2016. As part of the anniversary celebration, we initially thought of writing about all that we achieved and feel good about it. But this doesn’t help anyone, especially not the team. We need to realise how close we came to not scaling fast because of our choices and decisions.

Of the three blunders, two are big strategic decisions we as founders made and had we not corrected them in time, things would have been very different. The last blunder is a consequence of hyper-growth.

Young Internet startups are 1 wrong choice away from burning and crashing. In our case we did 3 but still managed to scale like crazy!

Hyperlocal: Hype Of 2015

This is definitely the biggest blunder we would have made had we stuck to this choice.

In early January, we started our operations in Mumbai working with some prominent VC-funded hyperlocal delivery startups. These partners gave us a phenomenal edge: deliveries in a few hours, scheduled deliveries, Fynd A Fit (try and buy at home) — everything that will make you skip the mall.

Why is this a mistake then? Multiple reasons:

Impossible To Scale

Operationally, the hyperlocal model is optimal to deliver for short distances: 3–5km. In our case, we want to deliver at a city level. Fashion stores are concentrated in malls and a few high streets. If you want to deliver across the city, then you need a lot of stores (which don’t exist). We realised this early on, and were able to manage till we had sub 100 orders per day by pushing our partners to do intra-city delivery. This also created an enormous engineering complexity on our order assignment logic.

The tipping point of getting rid of the hyperlocal strategy was a call with the kind folks of Zomato on April 29, 2016.

Limiting Our Customer Reach

Fynd is a real-time inventory exchange platform. Over 8,000+ stores sync close to real-time inventory with us. This information, when exposed in a hyperlocal setup severely limits the market size. Digital customer acquisition becomes extremely expensive. For instance, our Facebook CPI today is one-tenth of what it was when we were targeting only Mumbai.

Today we serve over 15,000+ pincodes and get orders from all over the country.

Evolved, simplified business model

Overnight Fleeing Of The Partners

Though this was not in our hands, post April 2016, one by one our hyperlocal partners started withdrawing their services. Initially, we thought it was limited to the smaller startups, but slowly even the bigger players disappeared. Imagine, as a company you are promising deliveries in a few hours and the next day you don’t have a partner to fulfill them.

Today we have partnered with large, well-established delivery partners who help us fulfill our customer promises.

App Only: What Was The Industry Smoking?

2015 was the year of the app. When we started building Fynd we met a lot of founders and investors to talk about our pivot from Shopsense to Fynd.

One thing we consistently heard was: make it app-only. No one visits websites anymore. Probably the worst myth of the industry.

We had never built a consumer app before and had little clue of Day 0 uninstalls, ASO Hacks, incentivised downloads, et al. Once we started understanding this better, we quickly understood what a phenomenally expensive strategy this is. It expensive to build, and prohibitively expensive to distribute (Facebook/Google discovery toll).

Fynd is not a daily habit app and people will install-uninstall based on natural buying patterns. Our purpose is to help people effortlessly discover their fashion, hence we will have to be on all platforms where people spend time — App, Desktop/Mobile Web and Messenger.

In hindsight, starting with the app is a good choice but staying app only is not. Starting with the app allowed us to build a highly differentiated app experience.

Hyper Hiring: Sloppy Team Building

Even though we have a unique lean model of building a store-led, inventory-less ecommerce startup, we still need a large team to build and manage multiple aspects — brand acquisition, engineering, fulfillment, customer support, growth, et al.

In early 2015, we were <20 people and by January 2016 we were 107. From early 2015 to August 2016 a total of 63 people left/were asked to leave. This constant flux of people leaving a startup placed lots of doubt on the wider team.

Bad strategy is easier to course correct. Bad hiring destroys your startup in front of your eyes.

Why did this happen? A few root causes:

1. Skipping the hiring process: We have a rigorous hiring process that was designed since the start of Shopsense. It works really well but takes time. When teams are asked to scale fast, shortcuts are taken.
2. Non-scalable Processes: This was largely because of following the hyperlocal strategy. We assumed some processes and created teams. These processes were operationally intensive and just not scalable. We have always maintained: anything that is not core, we need to find partners. This allows us to have a sharp execution focus.
3. Non-moldable legacy team
4. Consumer Internet Experts: The startup ecosystem is filled with marketing charlatans. We fired our marketing team twice! One of our biggest learnings is that Internet marketing experience is non-translatable. What works for one company doesn’t work for another — different time, context, newer channels, et al. It’s such a fast changing landscape that we need to do growth experiments using first principles and if they work, only then scale like crazy. Moreover, there are no silver bullets. What works once, does not always work again.

Today, as founders, we have absolute confidence in our model and have no doubt our team will push Fynd to steeper growth trajectories.

2016 has been a very tough year for the Indian Startup ecosystem. We at Fynd strongly believe in open sharing. Hopefully this article will help other startups, especially the ones in Bombay.

Note: We at Inc42 take our ethics very seriously. More information about it can be found here.