Recently, I was having a strategy meeting with a client as we were preparing for the upcoming year’s budgets. As our attention turned toward media budgets for the next year, and specifically to the digital media budget, the client stated that we should look at discontinuing all digital banner ads next year. When I asked why they would like to consider this move, I was surprised by the answer.
“We see very few to no conversions from our banner ads. I can’t go to my CEO and justify spending budget on banner ads with no ROI to show for it. We should put the money into other digital channels instead.”
First, I want to be very clear about my stance on banner ads. When you boil it all down, in most digital strategies, I like to use banner ads for three primary purposes:
- Provide top-of-funnel awareness for a brand, product, or promotion.
- Provide initial audience building to prepare for remarketing.
- Provide messaging for remarketing ad campaigns.
Banner as are incredibly useful as awareness tools. Traditionally, I don’t expect to report a lot of conversions from people clicking on banner ads. But banner ads still play a pivotal role in digital strategy. It’s all about how you measure the conversions.
Attribution models allow us to track and measure what media is most responsible for driving consumers to convert into new account holders or new customers. With digital marketing, attribution models break down not only what media is responsible for conversions, but which campaigns and even specific ads are responsible for conversions.
Depending on which attribution model is used will vary your conversion results. There are seven primary attribution models. Because marketing campaigns are not one-dimensional, we’ll use this example to explain the models:
John is shopping for a new car. While looking at car values, he comes across a banner ad for a credit union car loan. He clicks on the ad to check rates. A few days later, searching for loan rates, he clicks on a search ad for the credit union’s car loan. He visits the landing page and fills out a lead generation form for pre-approval. Next week, he receives an email campaign on the credit union car loan. The week after, before he goes out car shopping, he goes right to the credit union website and applies for the car loan.
- Last Interaction model: the entire conversion is attributed to the last touchpoint that the customer interacted with. In John’s case, it would be direct traffic to the website.
- Last Non-Direct model: in this model, all direct traffic is ignored. The entire conversion is attributed to the last marketing channel. In this case, the email John received.
- Last Advertising Click model: the entire conversion is attributed to the last paid advertising click. In John’s case, the search campaign would receive the conversion.
- First Interaction model: the entire conversion is attributed to the first touchpoint that the customer interacted with. In John’s scenario, the banner ad campaign would get the conversion.
- Linear model: each touchpoint the customer interacted with receives equal credit for the conversion. For this scenario, each would receive 25% of a single conversion.
- Time Decay model: each touchpoint receives a portion of the conversion. The touchpoints closest to the time of conversion receive more credit. The further in time away from the conversion, the less credit is received. In John’s case, the direct traffic would get the most credit and then the email. Less credit would be given to the search and display campaigns since they were over two weeks from the conversion.
- U-Shaped model: in this model, 40% of the conversion credit is assigned to the first and last touchpoints. The remaining 20% is evenly divided between any touchpoints in the middle. For John, the display ad and direct traffic would get 40% each. The search and email campaigns would split 20%.
In most cases, I prefer to use either the Linear or Time Decay model. This ensures that all touchpoints that a customer interacts with receive credit for the conversion.
Returning to my client story at the beginning of this post, there is a challenge when it comes to accurately measure and attribute conversions, even with an optimal attribution model in place.
Conversions require that a time-based conversion window is defined. Typically, most conversions will default to 30 days. That means only touchpoints that took place in the past 30 days will be counted toward the conversion itself. It helps in narrowing down what the most effective touchpoints are leading up to the sale.
In financial services, where a typical time-to-close for a conversion can take longer than 30 days, often those initial touchpoints will be left out of conversion attribution. Upping the conversion window past 30 days may help catch some of those touchpoints, but it will also dilute the value of the conversion as it attributes it across the channels.
In order to find out just how effective certain touchpoints of your digital marketing strategy are, without stretching your conversion window too much, there are a few things you can do to get the right information.
In the example from the beginning of this post, finding a way to measure how impactful the display campaigns are, beyond impressions and clicks, is the objective. With banner ads, measuring time spent on a page, or page views per session, gives insight into the effectiveness of the ad. If the site visitor simply bounces off the page after clicking the ad, we can render that visit not impactful.
This is where we might recommend a tactic such as setting up “mini conversion,” or conversions that consider page funnels and where a visitor might go from the landing page. Then, through AdWords or Analytics, comparisons can be made to all other paid media and alternate channels to find out how impactful display ads are at garnering the mini conversion.
Implementing solutions such as mini conversions and selecting the right attribution model are great starting points to going on to measure ROI on each channel. Happy converting!