Of Startups And Creating A Non-Scalable System : Why It’s Worth it

Paul Graham recently wrote an essay encouraging startups do things that don’t scale. His essay has got me thinking about one recent startup on how they approached the problem they were looking to solve.

The Problem: Rental classifieds on the web are only half useful as you still have to visit the location before you can make a decision. There are enough data fields for the user to add property info and pics but users don’t bother filling them. The data is either entered by an owner who does this once in couple of years, so its difficult to train him on best practices or by agents who are not always looking to rent out a particular property but are only interested in getting a contact made with the potential clients. As a result, the potential tenant is left with a half useful site that has most data fields empty and most filter options become meaningless.

Enter Housing.co.in – The great UX of this site would have fallen flat for just one reason – listings without pictures. The site would have been only as useful as incumbent players had they not had that the property pics. They have taken the pain of making sure every listing has pictures. From what I’ve learned, this wasn’t easy for Housing. They have about 60 employees in Bangalore only for engaging with brokers so that the data quality doesn’t suffer. I am sure that number is similar for other cities. It might look inefficient but hey, that’s their hare in the hat.

We have seen this happen in the past. When Flipkart thought of growing eCommerce faster, it set out to setup its own logistics system. At Freecharge we realized that retailers are yet to adapt to electronic coupons. Instead of hoping for them to change, we setup a whole system of distributing paper printed coupons. Even Google had people checking the quality of pages manually in the initial days.

Amazon had issues with clustering offers from different sellers under one product. The tool they made to solve that problem, Mechanical Turk, is now a product by itself. These steps are expensive and face a lot resistance from investors but with time, these become your biggest differentiator.

I am not sure what VCs mean when they say ‘it’s not scalable’. I don’t even know if that is a criticism or a compliment. What I do know is that it’s completely worth the pain and it’s not replicable over night.

[Guest article by Naman Sarawagi, Founder and CEO of FindYogi (has earlier worked with Zipdial and Freecharge).

Naman adds : At FindYogi, we are solving a small part of our problem in a manner that most people don’t approve of. We are collecting product data sheets manually. When we started, we believed that we could easily scrape content from ecommerce companies or sites like GSMArena but we were wrong. We still see that a lot of competitors and ecommerce companies who adopted to such practices have wrong/incomplete data. They will also change soon.]

Leave a Reply