Select Page

As you know from our prior post, Attribution Modeling Part 1: What Is It and Why Does It Matter, measuring your customers’ path to conversion can be a complicated proposition, one that grows in complexity as your online marketing program expands into new channels.  But Google—the company that has never met an analytical challenge it didn’t love and couldn’t conquer—has developed attribution models to help you more accurately understand where your customers are coming from and how they’re converting.

To better gauge which of these models will give the most accurate view of your channels’ influence, it’ll help to first review a few questions you asked yourself as you developed your marketing campaign.

  • Who are you? What is your business focus?
  • Who are your customers and what channels do they use? How are you reaching them and what do they want from you?
  • How do you measure success? Clicks, viral activity, transactions, subscriptions?

(You did ask and answer these questions, didn’t you? If not, call us; we really need to talk. You can’t reliably understand your channels and their success until you understand your business, your customers, and your image of its success.)

Keeping your key business identity and goals in mind, let’s look at a few quick examples that should help you zero in on what makes sense for your situation.

Small business owner or new to digital marketing — Use the Last Interaction Model 
As we described, the Last Interaction Model is the Google Analytics default: the last channel the customer touched gets credit. It’s a bit oversimplified, but if you’re small and/or just starting out with digital marketing—or if your business is geared towards quicker transactions instead of those that consumers mull over for a long time—it makes sense to test one or two channels and gauge their effectiveness. Get a bit more granular with variations on the last interaction model with, last non-direct click, last Adword click, which will help you remove the direct visit from being valued as the last interaction.  This “Last Interaction Model” will really help you hone in on the keywords and channels that really engage your customers.

Social media is your primary marketing channel — Use the First Interaction Model 
If you’re using Facebook and Twitter, then you’re probably most focused on getting the customer engaged, not necessarily pushing the hard sale. When you do have an actual conversion, you may overlook the channel that originally got your site on the customer’s radar screen. Using the First Interaction model helps show where social media’s working. (It’s also useful for initial branding campaigns and identifying the channels that first introduced consumers to your brand or new product.) Use the First Interaction model in tandem with Last Interaction for a Position Based model that gives weight to both the introduction and the closer.

Big Brand, or Big Campaign — Use the Linear Model 
If you have a well-known brand, with effective campaigns, you know effective campaigns, both digital and traditional. The Linear Model takes in all available information about your customer’s digital path and gives equal credit to all touch-points along the way. Though this model identifies what’s working and what needs tweaking, it doesn’t weigh the particular impact of any individual channel. If your customers move through many stages and channels along the way to conversion, the Linear Model might not capture what the real trigger is.

Your primary sales channel is direct marketing — Use the Model Comparison Tool Customization 
If you’re doing direct marketing, you’re experienced with the effectiveness of various channels. By creating Custom Credit Evaluations with the Model Comparison Tool, you can give more or less weight to particular channels (for example, your PPC click might be worth less than a click through from your blog). With a customized look at each channel, you can strengthen the parts of your strategy that aren’t working or fine tune those that are. This tool lets you get granular and rate both position and type of touch.

This should help you identify a model that seems right for you. Try it out, but keep in mind that every model has its pros and cons. The data you get can help you manage your digital marketing presence, but don’t just take it at face value: Question it, test and retest it, and verify it. As your business and customer base evolve and you revisit your online tactics, revisit your choice of attribution model as well.

One Last Thing…

So now you’ve got a grasp on Attribution Modeling and can readily analyze your digital marketing channels. Feels good, right? Well, hold that feeling for a moment. As a business operator you know your digital marketing is not the whole enchilada. You have other customer contact points at play: telephone sales, in-store interactions, even off-the-web application usage. Guess what? Google has some ideas on that and has rolled out Google Universal Analytics: a sort of uber-analytic tool that can integrate all your customer touches into one model.

But that’s a post for another day.