Choosing the Right Attribution Model for Your Paid Search Campaigns

April 18, 2019

Collective Measures
Imagine the last time you made a major purchase online. You probably spent some time searching around, browsing different websites, and clicking through various paid and organic search results until you narrowed in on exactly what you were looking for. Odds are you could have seen or clicked on several different search ads in the process.

Share this:

google ads attribution model

Imagine the last time you made a major purchase online. You probably spent some time searching around, browsing different websites, and clicking through various paid and organic search results until you narrowed in on exactly what you were looking for. Odds are you could have seen or clicked on several different search ads in the process. So, once you finally made a purchase and bought something, which ad (or touchpoint) was responsible for you buying that product from that brand? Was it the last ad you clicked before the purchase? Was it the first ad you clicked when you started researching options? Was it something in the middle?

In reality, it was probably a combination of everything you interacted with. Unfortunately, many advertisers use a last click attribution model to decide what touchpoint gets credit for a conversion, meaning the only touchpoint that matters is the last one a user clicks on before they convert. This model completely ignores most of the steps a user takes along the way, and generally isn’t an accurate representation of the real path to purchase that consumers take before they buy. Moving your paid search program away from last click attribution and to something that better represents the path your customers take to your product can help grow the overall health of your search marketing initiatives.

Users Are Clicking More Than One Paid Search Ad Before They Convert

The consumer journey doesn’t happen in a vacuum. Users interact with a brand at a variety of touchpoints across the internet on different devices and through different media channels before they convert. Paid search doesn’t happen in a vacuum either. Before a user makes a purchase, fills out a form, or takes a desired action on an advertiser’s website, they might search and click on four nonbrand keywords (keywords that relate to your products or services, but don’t include your company name), one branded keyword (a keyword that does include your company name) and then four more nonbrand searches before finally converting on a final branded query. Depending on the product, users can take weeks to convert.

Knowing all of this, how does it make sense to attribute the whole conversion to a single ad, keyword, or campaign when in reality that user clicked on several ads in different campaigns along the way? The short answer is that it doesn’t. Using a Google Ads non-last click attribution model can help make your whole paid search program more efficient. Not only will it give you better insight into which campaigns are assisting in conversions, but Google will use that information in your bidding strategies to optimize more effectively, and more in line with your ultimate business objectives.

Google Ads Attribution Model Options

There are several non-last click attribution models currently available in Google Ads. Depending on the attribution model, Google will “split” conversions and attribute them fractionally across different campaigns and keywords. For example, if a user clicks on an ad from a nonbrand query, then returns to click on an ad after searching a branded term and converts, each campaign would get a percentage of that conversion depending on the model you’re using. The currently available Google attribution models are:

paid search attribution model
  1. Last click: All credit is awarded to the last click before a conversion, regardless of the clicks that came before it.
  2. First click: Gives all of the credit to the first click, regardless of the length of the conversion path.
  3. Linear: Distributes credit for the conversion evenly across every click in a user’s path.
  4. Time decay: The closer the click is to the conversion, the more credit it gets.
  5. Position-based: Gives 40% of credit to both the first- and last-click and spreads the remaining 20% across all of the other clicks.
  6. Data-driven: Google’s sophisticated algorithm assigns fractional credit to each touchpoint based on past conversion data in an account.

Which Attribution Model Is Right For Me?

Every attribution model has pros and cons associated with it and each will change the way an advertiser optimizes their account. Some models tend to favor growth-oriented objectives while others favor efficiency, so it’s important to spend time closely examining each option to make sure you’re picking one that is aligned with your business goals.

Although for most advertisers a first click model is probably not the most effective attribution option, it may make sense for advertisers who are only interested in raw efficiency. If you are only bidding on a narrow range of branded keywords and want to drive as many conversions from those terms as possible, moving away from last click likely won’t have much impact. However, since many users begin their purchase journey on a nonbranded search and end up converting on a branded keyword, a last click attribution model can limit growth higher up the funnel by undervaluing anything that isn’t a branded search.

Like last click, the time decay model favors touchpoints lower in the funnel and undervalues higher funnel searches. It does, however, distribute some responsibility to clicks earlier in a user’s path and may be a good option for efficiency focused advertisers who still want to get a better picture of their customer’s path to purchase and begin optimizing towards searches higher up the funnel.

The linear and position-based models are a middle ground for advertisers who want to find a balance between growth and efficiency. Position-based heavily favors the first and last click a user makes with 40% of the credit going to both and distributes the remaining 20% evenly to all of the touchpoints in the middle. The linear model distributes credit evenly across every click, making no differentiation between clicks that occur earlier or later in a consumer’s path to purchase. However, advertisers should always take their business goals into account (growth vs efficiency) and remember that a middle of the road option may not make sense for everyone.

Far on the growth end of the spectrum, the first click model will be the most effective for advertisers interested in aggressive high funnel prospecting, seeking to drive new users to the site and expose new potential customers to their brand. Using a model like first click will drastically reduce efficiency however, since it will overvalue highly competitive upper funnel searches and undervalue low funnel branded keywords.

No attribution model is categorically better than any other, but the data-driven model has one major benefit that the others lack; It uses historical data from your specific account to decide which touchpoints are the most impactful and distributes conversion credit accordingly. This adds a layer of nuance that is absent from any of the other options. Data-driven attribution is only available to specific conversion points with enough conversion data however, so advertisers without enough account data may be limited to one of the other choices.

The most important thing to remember is that these models will influence what touchpoints and keywords you’re optimizing towards, so picking one that aligns with your goals and business objectives is critical.

Go Forth And Attribute

The key takeaway? There are several paid search attribution options to choose from, and last click may not be the best one for your business. Experimenting with different attribution models in both Google Ads and Google Analytics to see how different models attribute conversions to different campaigns and keywords is a great place to start. Once you’ve decided on one that makes the most sense for your business, switching your attribution model is as easy as going into your conversion settings and clicking a button.

Share this article

Share this: