Commentary
Defining Engagement Ain't That Hard
- by Cory Treffiletti , Featured Contributor, April 7, 2010
I don't get it.
Why is it so difficult for everyone in our industry to develop a measurement for engagement? Our business has so much data available; it's not a matter of finding the data, it's just a matter of finding the right data that you can work with.
The simple fact is, people are overcomplicating the issue. Engagement is an arbitrary term that the industry is looking to standardize -- but you can't standardize it because it's tied to individual brand objectives.
The broad term "engagement" refers to any metric that measures involvement, whether it is interactions, actions, time spent or other. It's a transitional metric that provides a bridge between exposure and end-actions like sales. Sales are typically the goal of any and all marketing, but sales data can be hard to come by, especially as an outside partner.
In the absence of that data, we rely on proxies. In traditional media we look at reach, frequency of exposure and brand metrics like awareness, consideration and intent. In digital media we look at all those same metrics, but we add the engagement metrics. The industry can create a bucket of metrics that fall under the umbrella of "engagement," but they cannot be standardized because each brand will be different.
advertisement
advertisement
I think it's actually a matter of laziness. Too many people are looking for someone else to solve the challenge for them. The challenge is, "how do we define engagement in a way that is accurate," by creating a true, relational proxy for sales.
You measure engagement in hopes of providing a close-to-real-time measure that you can optimize against, and that will provide an idea of the impact for your marketing on driving sales. The industry can't tell you what that is; you need to develop a formula and figure it out for yourself. You need to invest in analytics to understand the various data points that you have available and examine the correlations of these different data points over time. You develop an observation, you formulate a hypothesis and you either prove or disprove your ideas. The industry can provide guidance, but it can't do the work for you.
Engagement is a blanket term for individual metrics that you establish, test and utilize going forward. They are strategic in nature and they evolve over time as your business objectives evolve as well. The ideal situation is that you and your partners have a "data summit" to evaluate all the metrics you have available and begin to identify the correlations.
Maybe time spent on your site is relative to sales? Maybe interaction rates are relative to inbound inquiries, which is relative to sales? Maybe ad spend is relative to new visitors, which is relative to search volume on your site, which is relative to visits to the FAQ page, which is relative to sales? It can be a simple model or a complex model, but there is certainly a model that exists for your business.
Unless your business operates on pure Chaos Theory (which I truly doubt is the case, no matter what it feels like in the middle of the week), there is a model that will work. It just requires time, analysis and some more time.
Patience is not exactly my strongest virtue, but on this one issue I find peace, because I like the numbers. The numbers rarely lie, which is why I like media so much. The numbers can certainly be manipulated, but a good marketer can always cut to the important numbers, and those tell a story. They describe a path and they show you what your consumers are saying.
If you can hear what they are saying, and if you can see the path they are taking, then defining engagement based on the metrics they leave behind ain't gonna be that hard!
This post indicates that opinion makers are starting to embrace quantitative methods to not only describe, but also predict consumer behavior based on statistically valid models of engagement. To this end, I have worked with a consultancy called "Brand Keys" in NYC; they have a unique and successful methodology for measuring engagement across the entire range of media.
True story: One of the principals in a very popular agency where I worked modestly barked out to the media director in relations to a buy, "I don't care what the numbers say. Just as long as it works." It was a long time ago, but it left an impression.
Very thought provoking piece as usual, Cory.
But.
Spend five minutes in a supermarket.
100 kinds of cereal - 40 kinds of mustard - 70 cooking oils - all profitable - covering the spectrum of engagement from zero to ... whatever we geniuses eventually decide is a high score.
Think: most of the time people want the product, not the fabulously crafted "experience" that gets "added" to it and called valuable. Most makers of products will, I expect, eventually find themselves returning to this familiar and productive home base because they don't A) understand or B) believe in the fringe-y data-drenched but dollar-disconnected alternative.
A lot of what we call engagement these days seems to involve tarnishing or demolishing a brand, not building it. The hard truth is that thousands of brands/products fail a test of "engagement" but do just fine anyway, kind of like the ones that pass.
Measure or test every brand or product for engagement? The big problem is that some will fail (maybe yours!) and that engagement is not handcuffed to success, nor success to engagement.
Retail Media provides the opportunity to correlate offline sales with online engagement models due to the proximity of message delivery to the actual point of purchase. Can we provide a linear report illustrating the path beginning with a click and ending with an offline purchase - NO. What we can do is correlate that offline sales data (usually a spike) to impression delivery, clicks, etc. over a given time period. If you work on a Digital Team for any brand sold at Retail I urge you to take a good look at adding a Retail Media component to your plans whether it is a large or small piece of the puzzle. The fact is that your competitors are doing it and recent studies are showing that those embracing this "Digital Shopper Marketing" tactic are growing at a rate of 50% more than their category.
The measurement of 'success' is variable and objective...for ALL brands & products.
'Engagement' or 'Customer Touch Points' with Brands & Products need to be continually optimised over a period of time, so that LOYALTY is build between the consumer/customer/user and the brand.
Each consumer/customer/user is an individual and the data metrics on that persons behaviour need to be rationalised so that the initial objectives for that individual and his/her interactions with the offering are nurtured and supported.
Data metrics will show you the patterns - testing will tell you why the consumer/customer/user has selected the path to purchase (if the scenario is sales based).
I completely agree with your 'real-time' or 'close-to-real-time measurement' ethos. Real value in a retail proposition is asking someone 'Why did you buy that?'
If you really are interested, go and ask the people already engaging with your Brand & Products.
Engagement isn't difficult. The problem is humans love a complex explanation for things that are relatively simple. If it's too simple, we assume it can't be correct. Sadly, we are almost always wrong.
Engagement has several forms. But one of the purest ways of determining engagement is time spent on site. This is a double edged sword - if the user is engaged, do they care about the ads? I'd argue YES - because they are engaged, the recognize the values the site stands for and the offerings it provides are in line with what I enjoy.
On the other hand, not all time spent on site is equal. It takes a long time to read a news article, so news sites may have an inordinately long time on site - and while you're reading, you're not doing alot of interacting. Whereas if you're gaming, the ads are provided before a game starts....so you're already interacting. Or, if you're commenting on a site forum, you're visually aware of the content surrounding your commentary.
Another way of determining engagement is (shudder) click rates. I don't believe CTR is the be-all to end all of usage determination. It is like using a hammer to do the job of a scalpel. BUT it has its place in the toolbox.
When it comes to engagement, KISS remains the key thing to remember.