AB testing with call tracking

19 September 2014

Test everything. This is the modern marketers motto.

With a wealth of data available for those who know where to find it, AB testing has become a great way to prove hypotheses, test different messaging and even settle arguments in the workplace!

AB testing ads

Be it on search engines, TV ads, banner ads, indeed any ad at all, if the phone call is a call to action you want to measure then you can use call tracking to measure the success of these different ads. At the most basic level you could assign 2 static numbers to 2 different direct mail peices and see which letterbox drop generates the highest phone response. You can even tie phone calls back to each DM creative from visits to the website that were prompted by the direct mail piece if you use a vanity (unique) URL and dynamic numbers on your website.

Paid search is one area where many of our clients use call tracking to split test ad variations, keywords and ad headlines or copy. By using dynamic numbers on their website that the search ads point to, they can see which ad group, ad copy, keyword or campaign drive the highest amount of phone calls. All these insights populate Adwords, the other search platforms and even bid management tools so making decisions on the fly about the split tests results is easy.

Split Testing landing pages

Landing pages are commonly split tested for visits, web form completions, messaging, graphics, layout and many other elements – so why not phone calls? If you only measure one type of conversion then your landing pages will only be successful for that type of conversion. See the true effect that font changes, graphic and page layout decisions  and different call to actions have on every lead, phone calls included.

Here’s a example of how to run a simple A/B test for your landing pages and incorporate call tracking data in the results:

  1. Build your two landing pages with the different messaging, page layouts, and/or graphics you wish to test. Like with any AB test try to keep some consistent variables so the test is actually a valid one.
  2. Add dynamic numbers to your A and B landing pages.
  3. If you use Optimizely or another split testing tool that doesn’t give the A and B landing pages separate URLs then you’ll have to ask us to setup a separate pool of dynamic numbers for the B test. If you use Google’s split testing or any other tool that assigns the A and B tests unique URLs then you don’t need to make any changes to your dynamic numbers as Jet tracks the landing page URL people land on before calling.
  4. Ensure an equal split of traffic to each landing page across consistent times of days, geographies and devices to remove bias4. Determine how long your test will run. It should be long enough to collect a meaningful sample size of visitors.
  5. After the test is over, check Google Analytics or your analytics platform of choice and see which landing page generated the most web form completions and phone calls
  6. Pick a winner!

There’s plenty you can do with call tracking, our solution is flexible and robust so the limit is only an extension of your imagination.