Counting what counts

Publish date:
Social count:

Engagement. The most important, yet most poorly defined word in our industry. We know that’s what marketing is all about, but we’ve never agreed on what it is and how to measure it. Definitions can be expansive. Watching an ad on TV, talking to a salesperson at Target, sporting a brand on your hat, chatting with friends about products you like, diving into a branded digital experience — all are engagements and all matter, yet most of them happen outside of the view of any reliable measuring technology. Until we all have implants in our skulls, a lot of this just can’t be quantified.

Yet advertisers want to know what value they are getting for their money. Accountability is the word of the day. As it should be. But the language doesn’t exist yet to measure engagement

Image Title2

in all its forms. So the industry has stalled. It’s time to get moving again and to define engagement in terms of what is measurable.

That is only marginally less challenging. While the industry has grown, so too have the mechanisms that individual agencies, publishers, networks, and creative vendors have put in place to count what they deem to be important success metrics. And, to be honest, there’s a lot of game playing going on.

I could log onto a site, nick the corner of a banner with my mouse, have a massive expansion cover the content I really came to read, spend 15 seconds trying to figure out how that happened — and another trying to find the cleverly hidden close button — all the while cursing the site, the advertiser, and life in general. But somewhere an ad server is counting a 30-second engagement, and a media buyer is happily reporting success to the client. How many times have you seen an ad impression on a redirect page (hint: none — they’re invisible, but the server doesn’t know that)? Or an auto-playing video running silently beneath the scroll? Once, when I was on the agency side, I tried to figure out what great creative work had led to a super-high interaction rate, only to find that the vendor was counting clicks on the “close” button. And they were optimizing for the interaction rate. Imagine the expletives.

Even when everyone is being an upstanding citizen, we have a tower of Babel situation in the language of measurement. We have CPA, CPL, CTR, CPE, eCPM, complete views, brand lift, ad recall, time spent, and a host more. Agencies try to normalize all the different things they buy — and their unique measurement methodologies — through a Rosetta Stone of a spreadsheet that distills them into a common metric. But, naturally, every agency has a different spreadsheet.

It is crucial that, as an industry, we develop a common language to describe what constitutes engagement, perhaps with different categories of engagement (passive vs. active come to mind as very different groupings); what measurable indicators tell us that it’s happened and to what degree; and the effects we are able to quantify. Until we do, it will continue to be difficult to tell the C-suite that we are providing value.

We need to all be involved in the conversation and approach it with honesty, integrity, and respect for the ways consumers interact with advertising. What we need is to develop a measurement model that cuts through the murk and heads toward a vernacular that can compare all of these creative executions, technologies, and media types.

That model should have the following characteristics:

  • A trigger that validates the intent of the user. For VideoEgg, an engagement doesn’t start until a user has hovered over the ad for three consecutive seconds and waited through a 3-2-1 countdown. Is that the only way to qualify intent? Probably not. Show us a better one, and we’ll adopt it. But it’s intended to — and does — eliminate accidents.

  • A measure of the scale of the engagement. Time spent can be captured, and there is research that shows that greater time spent correlates to greater brand lift. But ideally, we would develop a system that accounts not only for time but for the various secondary actions that one can take with an ad. Clicks to a site? Great. Sharing seems great too, but how do we capture and evaluate the message that accompanies the share? Through research we need to determine the relative value of common secondary actions and normalize against these.

  • Post-exposure attribution tracking. Dynamic Logic is a common metric for brand lift, but because of the methodology — as well as the difficulty in retargeting engagers, as opposed to an audience exposed at the impression level — it is of limited value. View-through is a blunt instrument that assumes that a post-engagement brand interaction happens on a site the advertiser controls, which isn’t really the norm; if you’re considering a car, you might go to Tesla’s website or you might go to Until we have agreed-upon norms, every advertiser and agency needs to develop a model that works across both impression-based and engagement-based media.

It’s not an easy conversation for our industry to have. There are a lot of interests entrenched around certain ways of doing things. Just try to get different media vendors to agree on what factors constitute a qualified engagement

Image Title2

, and you’ll get answers ranging from clicks to rollovers to three-second hovers to who-knows-what-else. No one wants to change in a direction that will make their old metrics look like they were inflated — and that’s what’s likely to happen to almost everyone if we start describing engagement in terms of hand-raising, brand-recalling, time-spending, intention-peaking, action-taking consumer effects.

As an example, VideoEgg’s internal data shows that for every 21 rollovers on an ad unit, only one is intentional by our qualification methods. If that’s true generally — and data shared by some agency partners suggest that it’s in the ballpark — then there has been a lot of over-counting that will have to be adjusted. But if we can get over ourselves and our self-interests, we can do much better for the advertisers and, quite probably, learn a few things that will make consumers’ lives better.

The conversation only moves along if the following things happen:

  • Advertisers ask their agencies probing questions. How are they measuring engagement? Why did they choose to do it this way rather than another? How do they know it works? How do they normalize the measurement of different types of media? Don’t accept “industry standard practices” as an answer — those practices might not be the best for being standard.

  • Agencies ask probing questions of their vendors. How does each define an engagement? What consumer behaviors can we track? What behaviors should we track? How do you care for both the advertiser and the user experience? Are those being captured honestly?

  • Publishers ask probing questions of the agencies and the advertisers. Why do you think the metrics you have asked us to hit are the right measures? Can we avoid clichés and really talk about goals? How can we work more closely together to create value rather than talking about price all the time (if that’s the only measure that matters at the end of the day, we’ve all failed)?

  • The industry calls itself to action. This can be accomplished through conflict-free trade groups, by engaging in quality, controlled research, spending some real time and brainpower sorting through the issues raised above, and publishing a set of standards that enables understanding and eliminates game playing.

It’s not going to be easy. But if we’re going to find a better way to advertise in this very fragmented world that has outgrown its traditional metrics, it must be done.

By Matt Rosenberg is VP of solutions at VideoEgg.

Click here to see this article on