There’s an interesting billboard I see every time I drive between Dallas/Fort Worth and Austin, Texas. It reads, in colorful lettering: “Does advertising work? IT JUST DID!” It’s the filler message, put up when the billboard space hasn’t been sold, intended to entice potential advertisers to imagine their message in its place.
A similar question is often asked in the digital marketing space. “Is paid search incremental?” is one of those questions that all too often make us roll our eyes when it shows up in our inbox or on a phone call. Many digital marketers and agencies respond by trotting out canned stats or a prepared script, hoping they can reassure the client and avoid taking on a time-consuming project to prove the value of paid search.
There’s plenty of readily available data to back up either side of the debate. Google and Bing have released multiple studies showing that paid search has fantastic incremental value, and last year eBay released a study claiming the opposite.
Whichever conclusion you want to prove, there’s a source to support it – which is why it’s extremely important to test and determine the specific impact paid search has on your brand. iProspect has tested the incrementality of paid search many, many times – in fact, I’m currently collaborating with three separate teams to set up such tests for the near future. Here are three tips for determining the incremental value of paid search brings to your marketing mix.
1. Schedule the Test Thoughtfully
Ideally, it’s best to test any change to your marketing mix at a time when nothing else is changing – no sales, events, holidays, peak seasons, and so on. In the real world, that can be hard to coordinate. Still, you should endeavor to choose a low-volatility time to test. I’m a fan of scheduling such tests during your brand’s lightest months. For many retailers it’s often effective to test in June, after Mother’s Day and before the back-to-school season (unless you sell neckties, in which case you should probably wait until after Father’s Day). This has the added bonus of minimizing potential negative impact by concentrating the test outside of peak times.
It’s also crucial to plan enough time to gather accurate data. Dig into your site analytics and attribution data to identify the length of the conversion path, and make sure the test runs long enough to encompass at least one purchase cycle – ideally, more than one. Don’t forget to double-check your tracking windows to account for latent conversions that trickle in from paid search traffic immediately prior to the start of your test.
2. Establish a Baseline
I’m a fan of leveraging paid search geo-targeting to set up an A/B test. Identify a set of test and control markets with similar volume and customer behavior, and then pause paid search in the test markets. Then monitor performance in those markets as compared to the control markets. Make sure your site analytics and any other data sources are able to segment geographically and set up your report templates ahead of time.
Another approach is to compare year-over-year performance. This can be easier to execute, but if you choose this approach make sure to account for overall changes in your brand, your vertical, and the market from one year to the next.
A third approach is to split the test and control groups across search engines. If you’re running campaigns on both Google AdWords and Bing Ads (as you should be!), pause one and keep the other running, then examine the impact from each traffic source. As with geo-segmentation, before the test ensure your site analytics are properly configured to report on traffic from each.
3. Define “Incrementality”
This is the step that often gets missed. Look at each segment of your paid search campaigns (brand vs. non-brand, for example), and make a list of primary and secondary key performance indicators (KPIs). Paid search does far more than drive individual conversions. You should also be looking at new customer acquisition, new site visitors (to feed your remarketing lists), and secondary conversions such as phone calls. Hopefully you’re already tracking and reporting on these metrics – if not, this is a great time to start.
When analyzing the incrementality of paid search, take a look at each of these metrics. Track each for your baseline data and your test data and compare the two. Prioritize your key metric, but also consider the additional value paid search provides in order to get an accurate picture of the impact of turning it off.
Don’t forget to examine the performance of other channels during the test. The traffic lost from paid search in your test audience will either shift to your other marketing channels or be lost to competitors. Compare the lift in other channels (such as natural search) to the decrease in paid search for each metric to build your incrementality story.
I’m obviously a big believer in paid search – the real-time messaging in response to declared intent is valuable for almost every marketer. However, it’s important to quantify that value. The components that make paid search so successful (granularity, the ability to make quick changes, incredibly rich data) also make it one of the easiest channels to quantify in this fashion. Data from Google, Bing, and even eBay is useful – but every brand is different, and every audience is different. A carefully planned and executed incrementality test means you’re always ready to demonstrate the actual value of this channel to key stakeholders.