Digital marketing is measurable.
It’s probably the single most common claim everyone hears about digital, and I can’t count the number of times I’ve seen conference speakers talk about it (heck, I’ve even done it myself).
I mean, look at those offline dinosaurs, the argument goes. They all know that half their spend is wasted — they just don’t know which half.
Maybe the joke’s on us digital marketers though, who garnered only 41% of global ad spend even in 2017 after years of strong growth.
Unfortunately, while we were geeking out about attribution models and cross-device tracking, we were accidentally triggering a common human cognitive bias that kept us anchored on small amounts, leaving buckets of money on the table and fundamentally reducing our impact and access to the C-suite.
And what’s worse is that we have convinced ourselves that it’s a critical part of what makes digital marketing great. The simplest way to see this is to realize that, for most of us, I very much doubt that if you removed all our measurement ability we’d reduce our digital marketing investment to nothing.
In truth, of course, we’re nowhere close to measuring all the benefits of most of the things we do. We certainly track the last clicks, and we’re not bad at tracking any clicks on the path to conversion on the same device, but we generally suck at capturing:
- Anything that happens on a different device
- Brand awareness impacts that lead to much later improvements in conversion rate, average order value, or lifetime value
- Benefits of visibility or impressions that aren’t clicked
- Brand affinity generally
The cognitive bias that leads us astray
All of this means that the returns we report on tend to be just the most direct returns. This should be fine — it’s just a floor on the true value (“this activity has generated at least this much value for the brand”) — but the “anchoring” cognitive bias means that it messes with our minds and our clients’ minds. Anchoring is the process whereby we fixate on the first number we hear and subsequently estimate unknowns closer to the anchoring number than we should. Famous experiments have shown that even showing people a totally random number can drag their subsequent estimates up or down.
So even if the true value of our activity was 10x the measured value, we’d be stuck on estimating the true value as very close to the single concrete, exact number we heard along the way.
This tends to result in the measured value being seen as a ceiling on the true value. Other biases like the availability heuristic (which results in us overstating the likelihood of things that are easy to remember) tend to mean that we tend to want to factor in obvious ways that the direct value measurement could be overstating things, and leave to one side all the unmeasured extra value.
The mistake became a really big one because fortunately/unfortunately, the measured return in digital has often been enough to justify at least a reasonable level of the activity. If it hadn’t been (think the vanishingly small number of people who see a billboard and immediately buy a car within the next week when they weren’t otherwise going to do so) we’d have been forced to talk more about the other benefits. But we weren’t. So we lazily talked about the measured value, and about the measurability as a benefit and a differentiator.
The threats of relying on exact measurement
Not only do we leave a whole load of credit (read: cash) on the table, but it also leads to threats to measurability being seen as existential threats to digital marketing activity as a whole. We know that there are growing threats to measuring accurately, including regulatory, technological, and user-behavior shifts:
- GDPR and other privacy regulations are limiting what we are allowed to do (and, as platforms catch up, what we can do)
- Privacy features are being included in more products, added on by savvy consumers, or simply being set to be on by default more often, with even the biggest company in the world touting privacy as a core differentiator
- Users continue to increase the extent to which they research and buy across multiple devices
- Compared to early in Google’s rise, the lack of keyword-level analytics data and the rise of (not provided) means that we have far less visibility into the details than we used to when the narrative of measurability was being written
Now, imagine that the combination of these trends meant that you lost 100% of your analytics and data. Would it mean that your leads stopped? Would you immediately turn your website off? Stop marketing?
I suggest that the answer to all of that is “no.” There's a ton of value to digital marketing beyond the ability to track specific interactions.
We’re obviously not going to see our measurable insights disappear to zero, but for all the reasons I outlined above, it’s worth thinking about all the ways that our activities add value, how that value manifests, and some ways of proving it exists even if you can’t measure it.
How should we talk about value?
There are two pieces to the brand value puzzle:
- Figuring out the value of increasing brand awareness or affinity
- Understanding how our digital activities are changing said awareness or affinity
There's obviously a lot of research into brand valuations generally, and while it’s outside the scope of this piece to think about total brand value, it’s worth noting that some methodologies place as much as 75% of the enterprise value of even some large companies in the value of their brands:
Image source
My colleague Tom Capper has written about a variety of ways to measure changes in brand awareness, which attacks a good chunk of the second challenge. But challenge #1 remains: how do we figure out what it’s worth to carry out some marketing activity that changes brand awareness or affinity?
In a recent post, I discussed different ways of building marketing models and one of the methodologies I described might be useful for this - namely so-called “top-down” modelling which I defined as being about percentages and trends (as opposed to raw numbers and units of production).
The top-down approach
I’ve come up with two possible ways of modelling brand value in a transactional sense:
1. The Sherlock approach
“When you have eliminated the impossible, whatever remains, however improbable, must be the truth."
- Sherlock Holmes
The outline would be to take the total new revenue acquired in a period. Subtract from this any elements that can be attributed to specific acquisition channels; whatever remains must be brand. If this is in any way stable or predictable over multiple periods, you can use it as a baseline value from which to apply the methodologies outlined above for measuring changes in brand awareness and affinity.
2. Aggressive attribution
If you run normal first-touch attribution reports, the limitations of measurement (clearing cookies, multiple devices etc) mean that you will show first-touch revenue that seems somewhat implausible (e.g. email; email surely can’t be a first-touch source — how did they get on your email list in the first place?):
In this screenshot we see that although first-touch dramatically reduces the influence of direct, for instance, it still accounts for more than 15% of new revenue.
The aggressive attribution model takes total revenue and splits it between the acquisition channels (unbranded search, paid social, referral). A first pass on this would simply split it in the relative proportion to the size of each of those channels, effectively normalizing them, though you could build more sophisticated models.
Note that there is no way of perfectly identifying branded/unbranded organic search since (not provided) and so you’ll have to use a proxy like homepage search vs. non-homepage search.
But fundamentally, the argument here would be that any revenue coming from a “first touch” of:
- Branded search
- Direct
- Organic social
...was actually acquired previously via one of the acquisition channels and so we attempt to attribute it to those channels.
Even this under-represents brand value
Both of those methodologies are pretty aggressive — but they might still under-represent brand value. Here are two additional mechanics where brand drives organic search volume in ways I haven’t figured out how to measure yet:
Trusting Amazon to rank
I like reading on the Kindle. If I hear of a book I’d like to read, I’ll often Google the name of the book on its own and trust that Amazon will rank first or second so I can get to the Kindle page to buy it. This is effectively a branded search for Amazon (and if it doesn’t rank, I’ll likely follow up with a [book name amazon] search or head on over to Amazon to search there directly).
But because all I’ve appeared to do is search [book name] on Google and then click through to Amazon, there is nothing to differentiate this from an unbranded search.
Spotting brands you trust in the SERPs
I imagine we all have anecdotal experience of doing this: you do a search and you spot a website you know and trust (or where you have an account) ranking somewhere other than #1 and click on it regardless of position.
One time that I can specifically recall noticing this tendency growing in myself was when I started doing tons more baby-related searches after my first child was born. Up until that point, I had effectively zero brand affinity with anyone in the space, but I quickly grew to rate the content put out by babycentre (babycenter in the US) and I found myself often clicking on their result in position 3 or 4 even when I hadn’t set out to look for them, e.g. in results like this one:
It was fascinating to me to observe this behavior in myself because I had no real interaction with babycentre outside of search, and yet, by consistently ranking well across tons of long-tail queries and providing consistently good content and user experience I came to know and trust them and click on them even when they were outranked. I find this to be a great example because it is entirely self-contained within organic search. They built a brand effect through organic search and reaped the reward in increased organic search.
I have essentially no ideas on how to measure either of these effects. If you have any bright ideas, do let me know in the comments.
Budgets will come under pressure
My belief is that total digital budgets will continue to grow (especially as TV continues to fragment), but I also believe that individual budgets are going to come under scrutiny and pressure making this kind of thinking increasingly important.
We know that there is going to be pressure on referral traffic from Facebook following the recent news feed announcements, but there is also pressure on trust in Google:
- Before the recent news feed changes, slightly misleading stories had implied that Google had lost the top spot as the largest referrer of traffic (whereas in fact this was only briefly true in media)
- The growth of the mobile-first card view and richer and richer SERPs has led to declines in outbound CTR in some areas
- The increasingly black-box nature of Google’s algorithm and an increasing use of ML make the algorithm increasingly impenetrable and mean that we are having to do more testing on individual sites to understand what works
While I believe that the opportunity is large and still growing (see, for example, this slide showing Google growing as a referrer of traffic even as CTR has declined in some areas), it’s clear that the narrative is going to lead to more challenging conversations and budgets under increased scrutiny.
Can you justify your SEO investment?
What do you say when your CMO asks what you’re getting for your SEO investment?
What do you say when she asks whether the organic search opportunity is tapped out?
I’ll probably explore the answers to both these questions more in another post, but suffice it to say that I do a lot of thinking about these kinds of questions.
The first is why we have built our split-testing platform to make organic SEO investments measurable, quantifiable and accountable.
The second is why I think it’s super important to remember the big picture while the media is running around with their hair on fire. Media companies saw Facebook overtake Google as a traffic channel (and then are likely seeing that reverse right now), but most of the web has Google as the largest growing source of traffic and value.
The reality (from clickstream data) is that it's really easy to forget how long the long-tail is and how sparse search features and ads are on the extreme long-tail:
- Only 3–4% of all searches result in a click on an ad, for example. Google's incredible (and still growing) business is based on a small subset of commercial searches
- Google's share of all outbound referral traffic across the web is growing (and Facebook's is shrinking as they increasingly wall off their garden)
The opportunity is for smart brands to capitalize on a growing opportunity while their competitors sink time and money into a social space that is increasingly all about Facebook, and increasingly pay-to-play.
What do you think? Are you having these hard conversations with leadership? How are you measuring your digital brand’s value?
I have had a lot of these same thoughts for quite some time now, as I have clung to exact measurement and ROI calculations that are more than likely far from accurate. I've always just written my doubts off, or told people "there's additional value in how your brand grows, but I have no clue how to measure that for you." It's nice to know there are people out there thinking of ways to model this.
Love your comment on first-click attribution often not truly being first-click (or at least not first interaction).
Thanks so much Will, this was encouraging to read.
Important reminder that we can become walled-off in our own ideas. It's easy to forget history and ignore what happened in traditional advertising.
In the end it's all about the ROI. There's a reason companies still advertise cars on billboards, as the example used in the article. It might not be trackable, but it works.
Entirely agree with you Gil! Even though a lot of things move on to the digital world, does not mean traditional marketing is not as effective!
Heya Will! Long time, no talk. I had to stop by and leave a comment on your post because I've been feeling this way for a long time.
Back when I was working with Distilled as one of your favorite higher ed clients, we showed through a media mix model that our non-branded search advertising and Inbound marketing efforts (Thanks, Distilled!) were more responsible for driving brand awareness than were all of our brand advertising efforts combined.
It was seen as heresy, and I kind of enjoyed it... a lot :)
Also at the time, I was finding a lot of success in seeing how brand awareness and affinity changed by relying heavily on Google Consumer Surveys.
It was so effective that six years later, I find myself starting a company that uses consumer panels to measure brands. Brand marketing would become a lot less complicated if the industry had a "Google Analytics for brands," which is what we're trying to do.
Hey Josh. Great to hear from you!
I remember noises about that media mix model, but I don't think I ever got to see it myself. That's great work though.
I've wanted to do more with surveys tied to digital activity for a while - and my colleague wrote more about some of our thinking there (I buried a link to that in the post above).
Your startup sounds fascinating - and very much needed - I'd love beta access / to see what you're working on if you are able to share anything at this stage.
I totally noticed the article you linked to and immediately went over to Tom's post and commented. What we're doing is kind of like the survey approach Tom describes, but times ten. I'll hit you up on email for a personal tour :)
It's a very interesting article, the essence of human nature tends to lead to cognitive dissonance and confirmation bias, essentially we all think we are doing the right thing, or we wouldn't be doing it.
The clients we work with are often a big determining factor in how measurable our efforts can really be, the best ones allow you the time and money to sort out their measurement and reporting systems. We had a great client who allowed us to test everything they had been doing to date scientifically.
They had historically carried out a large number of newspaper adverts as well as digital and didn't know whether to continue press or not. We took what they'd been doing and reviewed branded mentions, direct website traffic (they sold B2C online only). We rebuilt all their press advert processes to include unique landing pages, phone numbers specifically for the campaigns, these did not deliver much and there was no real spike in the normal metrics. We tried various creatives, types of deals, long term voucher codes without much success. We stopped completely, expecting there to be an impact over time on direct traffic, but nothing changed!
We repeated the process with brand display advertising online, with similar results. When we ran through a similar testing process for paid search, it was a very different prospect. Although their customer purchases appeared to be largely price driven, they did have quite a long consideration period before purchase.
When we stopped the adverts that triggered for searches early on in the consideration process, the overall conversion for the same final click adverts dropped off dramatically, essentially seeing that advert early on reinforced, or put the brand in the customers mind for consideration. So even though lots of the testing (much more beside this) suggested they didn't have much brand equity and the customer was almost solely online, the brand awareness still had a large impact, but this was very difficult to measure.
It's not often that clients have the budget, or inclination to carry out such a large amount of work, so you have to use your best judgement a lot of the time. We carry out lots of work for smaller clients, they often carry out local magazine and flyer advertising, sometimes they think it works, other times they don't, which is why they come to us. We always try to test what they're doing before canning it, we often create a landing page for each advert, a unique offer, or setup some VOIP freephone numbers to see what's happening, the results are often surprising and not what either of us expect.
The main problem with all our marketing is we usually can't measure enough, even in our online digital world!
Digital might yet only have a percentage of advertising revenue, but the writing is on the wall.
And whilst I hope that the true brand value can be uncovered in time, it reminds me of that old joke when someone winds their car window down to ask for directions. "How do I get to higher profitability?, the guy rubs his chin and says "well I wouldn't start from here if I were you"... We are where we are, and as you say, it might only get worse.
There are many forces at work (in my head) around value and profitability.
I think about the tragedy of the marketing commons, and the related fact that many early projects will continue to show reductions in ROI because of it. Coming from the land of the Bleedingly Obvious ROI into having to working twice as hard to achieve the same results is less fun and 'challenging' dynamics.
How to counter..
Volume delivers more profitability, if the fixed costs are the same. And, Investments should increase over time, otherwise, we are just spending the interest gained and eating into growth. So, increasing the budget might be an answer.
A renewed focus on lower priced items - which are often more profitable than higher margin products with a longer buying cycle due to compound interest gained from those shorter investment cycles. Broadening the range is another way of getting more from the same traffic.
To sum up: getting less back, spending more, and focusing on lower priced products - all these seem boring compared to the good old days.
The only way to learn value seems to be to work out what an individual search term is worth - and that means using paid as some sort of divining tool... I can't see how else to do it - and then applying what we learn to the site. There is still so much time and effort wasted - probably in sales chasing unwanted leads - there must be room for improvement.
So if we do get a return visit - if we can link it to the first one, then we can attribute value (but only if we track forward to actual revenue, and not some guessed at average - which would only negate the result). Click to revenue feels like where it's at.
Sorry for the rambling reply, it seems that Moz always surfaces content that is 'of the moment' for me - some sort of global consciousness at work, thanks Will.
Interesting thoughts Matt.
Was wondering if you could expand a bit on this:
"A renewed focus on lower priced items - which are often more profitable than higher margin products with a longer buying cycle due to compound interest gained from those shorter investment cycles."
What would the compound interest be? A larger customer base? Shorter buying cycles leads to more purchases in a given time frame?
Really great perspective on this Will, we especially liked focusing on the value of increasing brand awareness or affinity. Thanks for writing and sharing your insight here Will!
So true this article is, we definitely have seen value outside of "shared metrics", especially long-term from out efforts in search optimization and digital marketing. Great article on clearing up this often overlooked misconception.
Consumer Journeys are complicated. In our experience, the hardest thing to track is when a company's CTA is a store walk-in.
1.The truth, I have often encountered difficulties, when it comes to investing in SEO.
SEO is not an instantaneous action, that's why it is very complicated to explain to a boss that he has to invest and the results are not guaranteed at 100%
There are 3 metrics SEO experts (contractors) need to prove (no other metrics matter AS MUCH)
There are 20+ other metrics that matter, but for C-level meetings, just go with these top 3.
In 15 years, none of these top 3 ever changed.
A great and timely article which I've sometimes struggled to explain client-side and agency-side. It's been obvious for years that just because it's 'digital' doesn't mean that it's completely measurable and 100% accurate for every possible value. And yet even after explaining it many times, clients are psychologically programmed to return to the default value in front of them every time...
And the idea that SEO is tapped out is certainly not true, so I'm looking forward to the follow up posts. I've been working with some clients recently who have seen brilliant results in an area bigger players seem to have ignored. And the overall volumes are growing in addition to their share of the market, so it looks even better in the future...
Very good, like all post here. The recognition of brand is very important. We have been applying it for a few months and it shows, people know us much more and are more interested. In fact, these last two months there are clients who look for us directly for the brand, although they include other factors, brand recognition has done a lot in this.
Regards!
Congratulations Jonathan! I know how hard is to create a brand in Spain and to start having recognition. Keep reading Moz and you will find success for your business; -)
Good luck!!
Hi Will,
Very good article thanks! Our agency LinkoPlus is struggling with that reality "which metrics to monitor and share to our clients". The truth is the clients want to see sales improvement and ultimately a better Return On Investment (ROI). For instance on Adwords, we focus on the ROAS and the ROI.
Regards,
Digital is only part of the game for us, and isn't right for everyone that we work with. We have been working to make sure that we don't just assume that this is the only solution to the problem, and while we invest a lot of time in it, there are definitely other areas we are really focusing on covering as well.