I came across a statistic a couple of weeks ago that intrigued me. According to a Journal of Marketing study, brands that invest in both social media and TV advertising sees a sales increase of 1% over a brand that invests in TV alone. And, the study notes, “A neutral or even negative social media post with high engagement will impact sales more than a positive post that draws no likes, comments, or shares. This is true even among customers who say their purchase decisions are not swayed by what they read on social media.”
In other words, customers who are more engaged are more likely to spend. And engagement is so contagious, that consumers only have to see that others are getting engaged to actually be swayed, whether they realize it or not.
That says a lot about the power of engagement.
Unfortunately, retailers and brands have long struggled to define “engagement” – what is it, and how do you know when you’re getting it? Is it something that can only be assessed in terms of “I’ll know it when I see it”? Does it really come down to nebulous things like clicks or likes or shares?
While social media activity does apparently lead to greater sales, as the Journal of Marketing study found, the connections are murky, and may not even be perceived as connections in the minds of consumers themselves. That makes measuring engagement tricky. And in the meantime, CMOs are finding less tolerance internally for spending built on “engagement” – like shares or likes – that don’t readily lead to sales, suggesting that retailers and brands are about to abandon investment in engagement in favor of activities that lead more directly to boosting sales.
The problem isn’t engagement itself, or even that marketing teams have sold their organizations on likes that didn’t lead to sales – the problem is with how retailers and brands measure engagement.
Retailers, with their more direct exposure to customer behavior than brands typically have, have a little more experience with customer engagement. They’ve tried to drive it in a couple of different ways in the past: Traffic, and more recently, Revenue Attribution. Both have their pros and cons.
Traffic is probably the oldest form of engagement measurement, outside of straight-up brand recall. Today, it’s one of the primary measures used to gauge the health of both stores and the website. If you never get people to come to your stores or your website, you’re never going to sell anything. Thus, traffic is a good indicator that people are aware of your brand to the point they’re willing to visit you – whether online or in-store.
What’s important about traffic is not the raw number, especially for stores. Here’s an example of what I mean. The least sophisticated way to measure store traffic is with a barrier counter, like infrared. You put it at every entrance, count every time the beam is broken, and divide by two to get an approximate count of people who came into and out of the store. The traffic number you get becomes the denominator in a conversion rate calculation: number of transactions divided by the number of people who came through the store. Retailers naturally want the conversion rate to be as high as possible, and of course store managers, with bonuses potentially on the line, will do a lot to make that number as small as possible, in order to make their conversion rate look better.
Here’s how that works. I once did some work with a consumer electronics retailer (that no longer exists) while that company was in the middle of wrangling with stores over traffic counting. The stores argued that having the traffic counters at knee-level counted “too many people” – because families came in as buying groups. That meant that even though the traffic counter counted four people walking into the store, there was really only one opportunity to convert to a sale – the odds of a family of four walking out with four TVs in one sale is pretty small. The retailer ended up raising the traffic counter to four feet high, as the compromise.
Another retailer, a fashion retailer, encountered a similar kind of pushback from stores in high-traffic tourists areas. They didn’t want to be held to the same kind of conversion rate expectation as lower-traffic suburban stores because the tourist stores got a lot of non-buying “entertainment” traffic, and large groups at a time. The stores argued their conversion rate target should be lower because they had a different kind of traffic.
But all of a sudden, we’ve moved from a discussion about traffic to one of how to compare stores against each other – and that’s part of the danger of using traffic. Traffic is a measure of engagement, because it basically tells you whether people are aware enough of your stores to go there. But that means the absolute number isn’t nearly as important as the trend. If your traffic is trending upwards, you’re doing something right when it comes to getting people into your stores. And once people are there, you have an opportunity to convert them.
If that number is trending down – as it has been for many retailers’ stores for a while now – then something is wrong. If retailers are missing the mark on sales, one of the first things they should be checking is how traffic is trending. The financial markets increasingly expect access to this kind of information from publicly traded retailers.
The same is true for online stores. People have to make it to the site if you expect to have a chance to sell to them, so at its most basic, traffic shows whether or not people are engaging with your brand.
One thing traffic does not do, though, is measure the quality of the engagement. Did consumers stay a long time on the site or hang out at the store? Or did they bounce right back out? Additional measures like bounce rates and dwell times are sometimes needed to get to some of that “quality of traffic” idea. And as the busy tourist stores pointed out, sometimes the makeup and buying intents of some kinds of traffic are different than others, so more sophisticated traffic tools – like video analytics – would be needed before retailers would be able to really draw any conclusions about the composition of traffic and the sales opportunities they represent.
But from an engagement standpoint, traffic trends are a good first indicator of whether or not consumers know about you – and care.
Revenue attribution is a more recent addition to engagement metrics, and it comes primarily from the online world. The challenge for retailers and brands investing across a lot of digital properties is in understanding where the investment worked – and where it didn’t. If a shopper first saw a Pinterest ad, and then a TV ad, and then happened to finally click on a banner ad to get to the retailer’s web site and buy something, how do you allocate that success across those three investments? Would she have clicked on the banner ad if she had not seen the TV ad? Would she have increased her awareness through the TV ad if she had not seen the Pinterest post first?
Revenue attribution attempts to solve this challenge through a set of complicated mathematics and behavioral analysis. And as more and more touchpoints become a part of customer engagement, the analysis becomes even more complex. If you want to burn some free time, you can always dive into the finer points of the philosophical arguments around first click vs. last click attribution.
The best analogy I’ve heard for revenue attribution is to liken it to proposing marriage. You can’t ask for marriage on the first date (well, you can but the success rate is likely to be very, very small). But you won’t ever get to the “last date” if you never had a first one. How do assign value to each step of the relationship, on whether it contributed to “yes I’ll marry you” at the end?
The easy answer is, you don’t. That’s awful. But on the other hand, which person do you think will have a higher likelihood of saying yes to a marriage proposal: the one who goes on dates only when you ask (and never asks you out), always stays for dinner but never drinks afterwards, and has never met your friends or family, or the person who asks you out, who enjoys spending time with your friends and family, who lingers over drinks…
One person is much more engaged than the other. Do you assign some fractional value to the drink purchased vs. the fruit salad brought to the family picnic? As if there is some equation that you could figure out would magically lead to “yes I will marry you” with every person you ask out on a date?
That’s the challenge with using revenue attribution-like methods to understand customer engagement. Creating engagement is much more about creating a relationship, and that means you can’t have a transaction – buy stuff! – as the goal for that relationship. The behavioral analysis behind revenue attribution can help measure engagement, and it serves a purpose for retailers trying to understand where to make digital marketing investments, but it is not a measure of engagement in and of itself – because it starts right off with a goal of “revenue” rather than “relationship”.
So What’s Left? Engagement Minutes Spent
Engagement, ultimately, is about time spent. It’s about mindshare. The one resource that everyone has, that can’t be bought or sold or traded, is time. So if a customer gives a brand some of her time, the brand should recognize that this is someone expending a valuable resource that can’t be renewed. This is engagement.
Measuring it is pretty simple: how much time does a consumer spend with you? Take a typical customer journey – how many touchpoints does an average consumer use? What’s the typical time spent per touchpoint? What’s the average time spent across those touchpoints?
Once you have a baseline, what’s most important is understanding how that time spent is trending. Is she spending more time or less? Is the typical shopper journey taking more time or less time? Are the touchpoints changing? Are there specific touchpoints that are seeing a significant change in the amount of time spent?
Then, and only then, should you look at sales – not as in conversion rate, but in the more general sense. Over a 30-60-90-day period, how much did a customer spend and how does that compare to the amount of time she spent with my brand? Is there a correlation to be made? Can you “rate” the quality of the engagement according to behavior – actions taken? Is a like or share a signal of greater engagement than just a view? (Probably!)
Once you have an understanding of trends, then it’s time to experiment, in order to determine if correlations are just coincident or indicators, or if they are actual levers that can influence engagement or sales. For example, if you find you have a higher level of sales from people who spend more time on average on Pinterest, if you invest to encourage shoppers to engage with you more on Pinterest, do you find that sales increase? If the answer is yes, then you have found a lever. If the answer is no, then you at least might want to monitor Pinterest engagement because if it starts trending down, you might have a leading indicator of falling customer engagement – which at some point or another will impact sales.
Once you know levers vs. indicators, you can start flipping the analysis to get more sophisticated. Start measuring:
- Total engagement time across all customers vs. total sales
- Digital vs. stores view – site engagement minutes vs. online sales. Store engagement/dwell time vs. store sales – how does it trend vs. overall?
- For high engagement/low sales: does that engagement feed other touchpoints?
- For low engagement/high sales: do shoppers come to that step via some other touchpoint?
The objective is to isolate touchpoints to get a better understanding of how, exactly, they influence customer engagement. For example:
- What is the engagement time spent for consumers who don’t use chat vs. those who do, and what’s the sales difference?
- What is the engagement time spent for consumers who talk to an employee (anywhere) vs. those who don’t, and what’s the sales difference?
The Bottom Line
Overall, the objectives for measuring engagement are simple. Understand how much time each shopper spends engaged with you in between buying events. Identify main shopping journeys and how long they take. Understand the trend over time of how engagement influences sales. Understand the role of each touchpoint in engagement, without getting hung up on whether that touchpoint specifically leads to a transaction. And use that understanding to invest appropriately to support strategic touchpoints that may not directly lead to sales, but certainly support the customer journey towards sales as the outcome.
Latest posts by Nikki Baird
- Retailers And Brands Measure Customer Engagement All Wrong - August 21, 2018