Credits to Tekkoontan

Had you had visited back in early 2014, you would’ve been greeted with the following message:

2 billion prices from 35 million products…”

The lead tagline on the front page of your website really sets the tone — and looking back, I can clearly see how inaccurate our product messaging and consequently, strategy were:

  1. Quantity vs Utility: Imagine if Google had marketed with the slogan — “we’ve crawled trillions of links”. That message does little to convey the true utility of Google. Our mission has been to power the future of data-driven ecommerce, but instead, we made it seem as though we’d merely gathered a huge pile of data.
  2. Selling FOMO: By selling customers on the wrong metric, we were effectively selling them on the Fear of Missing Out. The implicit pitch was that since we had so much data, customers simply had to sign-up or else they wouldn’t have access to a key strategic asset. Their competitors would have more data, or they’d miss out on that additional data point that’d make or break their business, and they’d fall behind the race in the era of data. Panic -> Buy.
  3. Importance of Size: The top 50 retailers in the world, barring a couple, cumulatively carry ~30 million distinct products. We were well past that number, and by the time we’d breached the 100 million mark, were entering the territory of 1 cent screws and custom printed tees. How much value were we really adding by pushing further on size?
  4. Non-standardized Metrics: These metrics were highly subject to interpretation. By some measures, we had 1 billion prices, and by others, we had 100 billion prices. We ourselves struggled to wrap our heads around these numbers. Imagine then, customers’ confusion.
  5. Internal Priorities: The focus on adding more data leeched away precious resources within the company from projects that optimized for data quality and customers needs (more on this later).

That wasn’t the worst of it though. When customers first sign-on, they tend to partially rely on vendors to understand where priorities should lie. By pushing customers to focus on the wrong metric, i.e., size, we were getting their projects off on the wrong footing. By the time reality had kicked in, precious time and resources had been wasted, customer happiness had dropped, and no one was content.

Something had to be done.

Back to the basics

We decided to go back to the basics — we started analyzing our API logs and conducted lengthy interviews with our customers to discover where our priorities should lie. We spent weeks talking to product companies and some of the top retailers in the country.

On size, we discovered was that the top 15 million products met 99% of customers’ needs. Customers rarely made use of the long-tail! They inevitably require just a sliver of relevant data for their business decisions

They needed us to find the needle in the haystack — instead, we were focused on expanding the barn and adding more hay to the pile.

What followed was a period of introspection, discussion and brainstorming (bookended by a team trip to Alaska during peak winter).

We emerged on the other side with a new set of priorities. Our business has changed in four key areas:

1. Responding to customer needs

In our discussions, turns out all of our customers prized three major features for their use cases. We decided to meet their demands head-on:

  1. Price freshness: In response, we introduced the RealTime API, the Push API and custom site refreshes.
  2. Long-tail access (when needed): In response, we introduced quick turnarounds on free site indexing of any ecommerce site of choice.
  3. Intuitive query tools: In response, we did a makeover of our API, culling distracting features and consolidating to a core set of queries.
Richer data. Fresher prices. Better API tools. Relevant results. Happiness.

2. Setting realistic customer expectations

We decided to be upfront and honest about our weaknesses. If it turns out we cannot support a customer’s needs, we let them know and no one wastes any time. Conversely, when we say we can get something done, we really mean it.

We like to think that we now guide customers to their purple patches rather than plunging them in an ocean of data and asking them to swim.

As a principle, we don’t consider a sale “won” if a customer doesn’t remain with us for at least 12 months

3. Achieving critical mass quickly

The more the data that we added, the fewer “Requests for Site Indexing” we received. The pattern followed that of an asymptotic graph, with the asymptote dependent not on the number of products, but what these products were.

This meant that our database is now dominated by demand-verified data, the best kind that you can have.

Demand is, it turns out, king. Now doesn’t that just seem frustratingly obvious?

We also did something completely counter-intuitive. We REMOVED data that was obviously useless, but help keep product counts and related vanity metrics high! Something we would’ve little considered 18 months ago.

4. “Infinite data” — break the walls

At the core, it’s always rankled us that our customer experience had to be limited to the walls of our database. In a world of unlimited everything, why were we even talking about finite sets?

To tackle this problem, our engineers developed the “Expanded UPC” feature.

When this feature is enabled, UPC requests are run not only across our curated database, but across the entire web(!!).

If ever there was a way to conclude an article that made the core dilemma moot, here you are!


All of this has panned out quite well for us. Particularly this year, we’ve seen a very fulfilling growth in customer happiness and consequently, sales.

Perhaps our greatest learning from this metamorphosis is that there are many ways to grow your business in the short-run.

Vanity metrics can inflate team ego and selling FOMO can make your top-line look good in the medium term, but neither of these makes for a sustainable business.

Focus on customer happiness though, and you will not go wrong! This, we’ve internalized, slowly but surely.

Are you looking for a external data solution for your ecommerce needs? Talk to us today!

Lovingly made in San Francisco, Singapore and Bengaluru, by Govind Chandrasekhar and the Semantics3 team