• by Mike May , May 29, 2013

Email drives sales. We know this already. How much sales it drives, however, is very often measured with a last-touch attribution model that has outlived its usefulness. If you’re relying on a last-touch model to measure your email’s effectiveness, you may be overstating its role in transactions. Equally as important, you may be understating the impact that some more aggressive email practices have on suppressing transactions, despite their high last-touch ROI.

The direct-response attribution model used most commonly in email today is little changed from the model pioneered by direct marketers half a century ago. The landscape difference between today and 1956 however, is that the Columbia Record Club marketed only through direct-response channels, with very little in the way of retweets and social shares to muddy the attribution. Today no brand exists in a single channel, so a 360-degree view of consumers is necessary to understand how each channel contributes to purchases. 

Channels characterized by branding and engagement certainly influence purchases – either by introducing consumers to the brand or greeting them with the right information at the right time or by helping to create an experience that ultimately leads to a purchase. But the last measurable touch before a purchase is rarely a customer service tweet or a :30 spot or some branded YouTube content. Unlike email, these ads are not direct response. This makes measuring their impact on purchases more challenging. But for marketers who invest in these channels, the effort and intelligence are necessary in order to allocate some percentage of revenues into the ROI calculations for social, advertising, events and sponsorship.

Of course, recognizing that other channels have a legitimate claim to some sales that email touched last means that the “R” in email’s ROI decreases. Relinquishing some of the revenue attributed to email in a last-touch model may be an unpopular suggestion to some, but it’s the only route to an informed conversation about multichannel marketing. The Direct Marketing Association, which publishes an annual study about ROI, found most recently that every $1 spent in email drives about $40 in sales. Email champions positively love data like that, and I admit I’ve touted it myself. But in modern marketing, measuring each channel in a silo misrepresents how brands actually drive sales. It is doubtful that ROI would still hold if the brand ceased all TV advertising, yanked its search spending and shut down its social presence.

What really drives revenues in modern marketing is no single channel, but the overall user experience comprised of brand interactions across all channels. What a multichannel attribution model does is figure out how much each channel should shape the user experience for the greatest impact on sales.

When you look at sales in the context of user experience, it is clear that what the user experience giveth, the user experience can taketh away. Exceptional UX drives sales. A poor UX stultifies them. For example, a well-timed triggered message, after a delightful online shopping experience precipitated by a retargeted banner that stemmed from a Google search, prompted by a TV ad, might seal the deal on an abandoned shopping cart. But if that triggered message begins, “Hi Nancy!” when the shopper’s name is Carl, Carl may well suspect the brand is not careful with the data he has entrusted them with, and leave his cart abandoned forever.

As email marketers, we often seek credit for email’s role in driving offline and latent online transactions. We should be equally ready to take responsibility when overzealous sending and inaccurate data management repel consumers, compromising both online and offline transactions.

If you’re hoping to find here the handy bullet-point list of how to build a multichannel attribution model in five easy steps, I’m afraid I’m going to disappoint you. A model of any accuracy or sophistication is extraordinarily complex, and for many brands true multichannel attribution is simply unknowable. But that doesn’t mean that last-touch is the next best alternative, particularly if it perpetuates practices like overmailing and metrics-chasing that ultimately undermine the user experience -- and suppress sales not measured as closely as the ones that do occur immediately after clicking a link in the inbox. Sometimes we have to make strategic decisions based on our brand values and in the absence of perfect data  -- which, by the way, ends up being a pretty good working definition of “strategic decision.”