The eCommerce Search Conundrum: Site Search’s Role in the Marketing Platform

Last week at the Internet Retailer Conference, I noticed an awful lot of search vendors on the expo floor. Not just the big players who have been around for a while, but a lot of little startups too. Why this influx of site search capabilities into the market? And how sustainable is it to have this many players long term?

In retail stores, there was a study in the early 2000′s that showed that if consumers can’t find an item they’re looking for in the store, they assume that the retailer is out of stock. Some doctoral students did some math on measuring how much this perception impacts the retailer, and they found that a retailer whose store may be at a 90% service level against the inventory it expects to have, could find itself at only a 70% perceived service level – because consumers couldn’t find the items they were looking for and thus assumed that the items weren’t there.

For eCommerce, site search serves much the same function. In fact, it’s even more important, because the effective “sales floor” of a site is much smaller than the sales floor of a store. If I go to Home Depot looking for a PVC pipe fitting, there might be two or even three places I could look for it in the store. I know that, they know that, and they do their best to help me find it between signage and the now-ubiquitous orange-aproned people walking the floor.



On a site, if a shopper doesn’t find what they’re looking for by browsing, they turn to search (some people skip the browse all together). And if what they’re looking for doesn’t come up on page one or two of the results, it’s like the retailer never had it. That’s a far smaller percentage of products viewed in that search than the same kind of search conducted in a store – and a far larger chance of the retailer to end up missing the sale.

So search is critically important to sites and to making sure that retailers get the best chance of putting the right inventory in front of the right shopper. But then why is there this sudden rash of search solutions? It all comes down to the role search plays in the overall marketing platform – and sorting through the pieces and parts of an end-to-end customer marketing platform is opening up a giant can of worms. In trying to sort through all that, here’s what I’ve come up with:

At the heart of search is a customer profile. Search looks at a shopper’s behavior and builds a profile over time. Really smart search capabilities are able to compare a profile to other profiles and start anticipating patterns of behavior based on similarities to other behaviors it has seen in the past. For search, that is expressed as learning things like people who search for “bathroom cabinets” are looking for the same thing as people searching for “vanities”.

That sounds a lot like personalization, and it should – search and personalization are increasingly intertwined. This is because once you have a shopper profile, you want to be able to do something with it. The question retailers have to answer is, “Is this customer profile merely A customer profile, or is it THE customer profile?” Is it the profile that collects and correlates all of the information that a retailer has about a shopper? Retailers have already strongly expressed a preference for as few customer profiles as they can manage. They don’t want one for search, one for loyalty, one for social media listening, one for store behavior, etc. They want one customer profile that tracks everything.

So search alone is a feature, but it’s a feature built on a core capability that could become much more central to a retailer’s marketing operations.

Where startups proliferate in search is around that second half – what they do with that customer profile – the algorithms that make recommendations based on past behavior or the behavior of other similar profiles. So many startups in search seem to be competing on the idea that their math is smarter and better than the math you already have, which is an extremely challenging way to compete. Especially when “search” alone, while critically important, is also a one-hit wonder.

Search alone doesn’t have enough gravity to be a stand-alone application, certainly not for very long. Between the commerce platform vendors who could just beef up their search, and the more mature search vendors who are already developing expanded capabilities that aim to make them THE owner of the customer profile, “better math” probably doesn’t stand much chance of winning in the marketplace.

I overheard a colleague from a past life once describe the challenge that companies face in selecting software like this: “Retailers have a list of 10 things they want. And they look out in the marketplace and find that all the vendors have 5-7 of those things, and none of those vendors are mutually exclusive – there’s a lot of overlap. So a retailer might have to get four solutions, of which 25% of those solutions overlap, just to get their list of 10.”

For retailers looking at search, personalization, and marketing platforms, the choice is the same. The most important question to ask is not whose math is better. It is, “Where should my one customer profile live?” Does it live in a marketing platform? A loyalty or CRM solution? And how connected to action does it need to be – a customer profile in search turns insights into action almost immediately. A customer profile in a CRM solution might struggle to do the same.

A conundrum indeed.