• by Cabell De Marcellus , Op-Ed Contributor, July 6, 2017

With 90% of advertisers saying that they’re reviewing their programmatic advertising contracts and demanding more control and transparency, you know there's a transparency problem. In programmatic, there are brand safety and measurement issues because too many advertisers have too little insight into where their ads are running and how each site is truly performing -- e.g., is that traffic legit, or is it from a bot?

Demanding more transparency from agencies and platforms is one way to address the problem. Advertisers can also take steps on their own to track the effectiveness of a campaign. It starts with knowing on what publishers their ads are running.

Armed with this information, they can begin to effectively understand how their ad dollars are performing and to proactively measure and optimize their campaigns. Here are a few ways to do so:

-- Use Google Analytics UTM codes.  A UTM code is a tag that marketers can use to track a source (such as a publisher) and campaign. By adding a UTM tag to a campaign URL, they can use Google Analytics to see what source (publisher) the traffic came from and what campaign directed them. This will allow them to see all their activity by publisher, including: clicks, ROI metrics, CPA (cost per acquisition), device and time on site.  

-- Optimize by device. Using UTM codes will help you narrow in on how individual publishers are performing. You should also understand how different devices are performing. Many platforms allow you to select which devices to run ads on: desktop, tablet and/or mobile. By differentiating campaigns by device, you’ll get transparency for the clicks.

-- AB Test.  Make sure to A/B/n test your ad variants. This means running two or more versions simultaneously and measuring for clicks and actions. Be sure to collect enough data so the results are statistically significant.

There are many free tools online where you can input the data into and get a P-level, the probability that the result is due to random error. You should look for a P-level of under 0.05 or 5%. Try http://getdatadriven.com/ab-significance-test.

Advertisers should look closely at traffic from each publisher, analyzing it over time to compile a holistic understanding on their own of how their campaigns are performing against each site and device. Armed with this data, they can have more meaningful and proactive conversations with their agencies about ad buys, and can dig deeper and flag sudden changes such as an upsurge in traffic that could indicate a publisher is buying traffic or is a victim of click fraud.