American Marketer


Stop failing at mobile-campaign attribution

April 24, 2015

BJ Hatcher BJ Hatcher


By BJ Hatcher

How many of your mobile campaigns simply fail to generate the metrics that you really need to justify your mobile ad budget? I am guessing more than your directors would prefer.

Do you use the excuse that your campaigns are driving more sales but you just cannot measure them? Are your directors weary of hearing that over and over again? I am guessing so. But there are better ways.

Campaigning for
Attribution is a pain. Yet it is fundamentally the most important part of any campaign that you run.

There are a multitude of offerings in the mobile marketing space that have made great strides in the attribution of offline behavior at its core. They all range in cost, time needed and overall complexity.

I am going to focus primarily on CRM data-driven marketing here.

List-driven digital campaigns are a direct marketer’s dream come true in the digital ad space – and even more critical now that they are taking a stronger foothold with a wider net of mobile networks.

If you have a set of individuals or even just email addresses from your database, marketers can match those records online – both mobile and Web – and serve an ad campaign focused to those individuals.

I acknowledge that this is counter to what many media partners may recommend, yet it is very intentional that brands should take this approach to campaigns.

Marketers should not care about optimizing for clicks, because clicks do not help your business – sales do. And unless the bulk of your transactions happen on your Web site, you need to have a better way of understanding those offline transactions.

The goal is to replace more expensive tactics with less expensive digital placements without sacrificing performance at all. More bang. More buck.

Listing reasons
Marketers should generally start with a targeted list of consumers and match that list to the same number of digital shadows across device types. Then serve standard, rich or video banners to those devices – and only those devices.

Avoid additional propensity models to find more people like the targeted list, as that original list is all you are focused on. Why belabor this point? Because it is important.

For the modeled campaigns, unless you bring in additional attribution partners such as VisualIQ, Adometry or Convertro, once those modeled recipients come into your stores, you no longer have a starting list to measure them against. So you are back to making up excuses about why your campaigns do not work in the first place. Stick to your list.

The campaign is live, and sales are coming in. And because the marketer stuck to its list, it can now watch for those individuals to purchase.

Plus, with the addition of a non-treated control group, you can measure the overall impact of your treated banner group to the non-treated group and claim sales – across your entire organization.

NOW YOU can clearly make this statement about your mobile campaigns: We targeted 1 million individuals and generated an additional 10,000 sales across in-store, call center and online channels when compared to non-treated groups. That sure as hell beats all of the excuses you have been telling your boss.

It really is a simple process at the surface and no different from how mailings, direct mail or email, are executed for your marketing programs.

However, there is a whole host of complexity just below the surface that is worth discussing.

BJ Hatcher is account director at Hacker Agency, a Seattle-based digital and direct marketing agency. Reach him at