LIST AND DATA STRATEGIES caught up recently with Chris McDonald, executive vice president of cooperative database Abacus and data management for Broomfield, CO-based database marketing services provider DoubleClick.
LDS: What is the state of cooperative databases today? Is there too much saturation?
McDonald: What we’re seeing in the marketplace is that three [co-op databases] entered the market at a similar point and with a similar concept around matching SKU-level data. SKU-level data was thought to be the next evolution of how to source names in the market, and what we’re now seeing is that the concept hasn’t really become mainstream. They are finding very small incremental universes of names.
I think the conclusion you can draw today is there still are not any materially more new names on the market than there were three to five years ago.
LDS: What’s the next wave in the marketplace?
McDonald: We decided to invest in multichannel [data]. The key reason is if you go back and organize your data by channel and see the interaction between those channels we can create higher-performing prospect models.
Suppose a customer goes and hits Google and types in “Ionic Breeze,” and it takes them to the Sharper Image Website, and he can buy a refurbished Ionic Breeze at 50%-70% off the market price. If that buyer responds and makes a purchase, when [the data come] into the Abacus Alliance, is that buyer going to be representative of someone we should model from to find more buyers? Absolutely not. In organizing data by channels we understand which buyers you want to bring in to create an affinity model because you want as pure a model of who your best customers are as possible. Because if you have people coming in who aren’t your best customers, who are price-sensitive one-time buyers, your models are going to have some “noise” in them and not respond as well.
LDS: What are some of the biggest concerns coming at you from clients?
McDonald: Clients are starting to hold co-ops more accountable. We’re seeing clients make other co-ops now be incremental to Abacus, because a lot of the names the other co-ops were shipping we already provide. So you go into a merge and you start allocating names that you are already getting from other lists, and it disguises how good those other lists really are. We still net, in some cases, 60%, 70%, and 80% unique [names]…so now people are starting to say, “I need to reduce the number of co-ops I work with, and I need to hold those co-ops accountable.”