Welcome to The Combine — the official TellApart blog. With the unveiling of our new website you’re probably wondering, what’s up with all the wheat?
When we were Entrepreneurs in Residence at Greylock Partners, we started working on a bunch of ideas which centered around big data problems. But, as is true with most data sets, the size of the corpus alone is not what’s important… it’s the ability to identify the meaningful features within the data, and then to build the feedback loops on top, that let us refine the big data into big value. And so, we nicknamed our efforts “Project Combine” in a nod to the software machinery we would build to help separate this digital wheat from the chaff.
As we dived into the most pressing problems of our e-commerce clients over the past year, we immediately identified a common thread. While these companies were being bombarded with marketing pitches from companies trying to sell them new data (both online and offline), their own customer data was hardly being used. Basic problems pervaded: customer data stored at the transaction level instead of being rolled up to global customer IDs, transactional data entirely and irrevocably separated from web analytics data, no way to determine if a current website visitor had ever been an earlier buyer, etc.
For decades, offline direct marketers (catalogers, loyalty card programs, casinos, etc.) have invested stunning amounts of effort and money in the management of their customer data. Catalogers, especially, knew that profit and revenue growth were found in the careful manicuring of their “house file” — a comprehensive list of everyone who had been mailed and who had made a past purchase, down to exactly when and what they bought. The predictive strength of this data, along with the modeling and data cooperatives commonly used in the industry, meant that every person could be given a score representing their likelihood of purchasing again. And this score was used to determine who would receive a given mailing — either as a repeat or entirely new customer.
So why is it that many of these lessons have yet to be applied online? Because most of the industry’s efforts for the past eight years have gone toward optimizing the search funnel. Now that search marketing has become tremendously expensive and has reached a point of diminishing returns, we believe the time has come for e-commerce companies to return focus to their customer base. We founded TellApart to adapt and transfer this knowledge.
In addition to getting this data out from behind the firewall, syncing it in format that can be unified at the customer level and putting it in the cloud, companies need to make this information addressable. It’s no longer about the names and addresses in the house file… it’s about the user list: virtual representations of customers backed by tremendous amounts of online data — which we model to help our clients “tell apart” their highest value users from the rest. Simply, the cookie list is the new mailing list. Once identified, these high value customers can begin to have differentiated experiences — both on and off the e-commerce site — through channels like display ads, offers, recommendations and improved customer service. We believe Customer Quality can be measured and is the cornerstone to refining all marketing efforts in the future.
Our team includes some of the best minds from online retail, finance, and software companies like Google, Oracle and LinkedIn. And we have collective deep knowledge and experience in the areas of high performance cloud-based computing, massive scale data storage frameworks like Hadoop, e-commerce analytics and online advertising.
If the caliber of these people and these kind of ideas excite you, subscribe to our feed, follow us on Twitter, contact us below, or check out our open positions! We look forward to sharing our progress in building TellApart into the de-facto customer data platform for online commerce companies.
Josh McFarland & Mark Ayzenshtat — Founders, TellApart