February 19, 2018, posted by shahzeb arif

Which Numbers Count? Understanding Paid Media Analytics

Every action we take on a digital device is measurable, hopefully anonymously, but measurable nevertheless. There has always been a need to measure the performance of advertising, which produced Nielson ratings for TV, Arbitron for radio, and the Audit Bureau of Circulation for print. Digital measurability, however, has restructured our expectations to believe we should be able to fully understand the effect of every marketing dollar we spend. While we are closer, this is not yet true, and measurement remains a combination of art and science.

The science is in measuring activity. We can and do that with everything, especially digital marketing. The art is in picking the right measures to pay attention to and then drawing the correct conclusions. Setting KPIs (key performance indicators,) which are important to the business, is essential. Sales, of course, is the ultimate KPI, but connecting marketing metrics to off-line sales is still a trick. Invariably there is a tracking gap between marketing activity, and the sale itself. This can misrepresent the influence of any particular channel on the sale.

In paid digital advertising there are numerous metrics. A sampling includes impressions or cost per thousand (CPM), unique visitors, click through rate (CTR), cost per click (CPC), cost per lead (CPL), percentage complete, and time-spent viewing, etc. The idea is to figure out the type of behavior that gets closest to inferring the outcome you want from the campaign. For example, a pure brand campaign, designed to create awareness and brand favorability, might focus on the time a prospect spent viewing a landing page, which they arrived at by clicking an ad. This would tell you the degree to which the prospect was interested and engaged, which we might infer would increase brand awareness and favorability. While these metrics would not directly measure sales or intent to purchase, they would give us an indication of the effectiveness of our brand engagement effort.

Advertising and marketing today, however, are not a series of isolated activities. Instead, they are part of a marketing value chain, in which many consumer touchpoints layer on and interact with each other until the consumer is ready to act. For that reason, the KPIs we select for any tactic should reflect the step in the consumer journey that the campaign is addressing. So, while a brand awareness campaign, at the beginning of the journey, would measure engagement, a demand generation campaign, later in the journey, would measure conversion.

A growing solution to tracking this is Multi-touch Attribution. This is the methodology of assigning weighted credit to the ad touches which lead to a conversion. How important was the last touch? How did previous touches play a part? Although still very imperfect it will be very influential in the future and is preferable to only giving credit to the last thing someone sees. It may also give us new insights into the layering effects of multi-touch marketing, which goes beyond ads to content and digital.

Teams should produce weekly campaign reports that not only show activity metrics, but also performance against KPI goals. These are often best integrated into dashboards so that business leaders can see the greater meaning of consolidated data across all marketing activity. Traditional research is also still a valuable tool to determine important psychological measures such as brand recall or purchase intent, which are difficult to measure with activity metrics. Once again, these insights are still a couple of steps from connecting to sales but get you closer. If your target audience shows solid brand awareness, brand preference and intent to purchase, you know you are on track to get conversions.

Resist the temptation to look at sales for the period figuring that if they go up the marketing worked, and if they go down it didn’t. The problem is that while there is probably truth in this, it does not enable a marketing organization to become smarter and more effective over time by learning. The only way to do that is to know which parts of your marketing are working and why, so you can constantly improve your performance.

The ability to adjust media campaigns on the fly is called optimization. The concept is fairly straight-forwarded: take the current performance of different advertising vehicles, evaluate them regarding your goals, then cut those that perform poorly and heavy-up on those that are doing well; it’s sort of like rebalancing your stock portfolio. Studying performance tells you not only which network or website to put your ads on, but also often what period to use, what position to buy, and even which ad creative to run. The idea is to identify the best performing vehicle, context, and message, and then funnel your dollars there.

The beginning part of any media campaign is the learning phase, where you are actively testing your best guesses, based on your strategic planning, for what the right advertising channels are and what the right messaging will be. In a perfect world, you will have researched before launching the campaign to make sure your choices are right, but it’s almost never a perfect world. Even with research, it’s impossible to fully know what’s going to work until you try it, so your campaign should plan to test alternative media and creative approaches right from the outset. This is often easier with media than with creative, since making multiple video commercials, for example, is often less financially feasible. As part of your buying process with media channels you should negotiate complete flexibility, so you can quickly react to the data you see. Many platforms build in the capability for multivariate testing, and the ability to make fast changes to media buys. This can make a big difference to the final performance of a campaign, so make sure that your team has a plan for optimization and testing for all your purchased advertising.

With the insights and strategies that you have developed as part of your marketing plan, plus a strong, healthy brand, you are ready to introduce your brand to your prospects. This is where paid media comes in and requires a general understanding of metrics, analytics, and optimization.

Share on Facebook Share on Twitter Share on LinkedIn

Measurement What You Don’t Know Will Kill You

Measurement Putting Your Customer Data to Work

Measurement 3 Questions Predictive Modeling Can Answer for Marketers

Advertising, Digital First Entertainment: Young Storytellers

Advertising, Branding Pergo: Floors That Don’t Judge

Advertising, Digital Daltile: Rewrite The House Rules