Ahhhh, the “vanity metric”: that all-purpose filler of pitch-decks, scourge of data analysts, and lifeblood of Internet marketing charlatans everywhere.
Countless articles have been written warning of the danger they pose to organizational decision-making. Untold volumes of digital ink have been spilled urging marketers to focus less on the numbers that make them feel good about their work, and more on the numbers that tell them whether or not their work is good.
And yet, so much of the history of online marketing is the story of fortunes won – and lost – by appealing to marketers’ collective sense of professional vanity.
We saw this in the days when SEO firms guaranteed that their clients would “rank #1 on Google”, and when social media gurus promised “viral success” in hazy, nebulous terms. Even now, our industry is caught up in a debate over the value and merits of “influencer marketing” – a subject on which I’ve made my opinion fairly clear.
It seems like every couple of years, another new sales tactic comes along which appeals directly to our industry’s love for vanity metrics, and the whole sorry cycle repeats itself again. The ‘winners’ in this process have typically been those aforementioned charlatans, and the ‘losers’, their clientele.
Whenever the topic of vanity metrics comes up, it’s usually those “top-of-funnel” numbers – pageviews to a website, likes on your social media posts, total downloads for an app – which get singled out as the main offenders. Which makes sense, really: these are often the biggest numbers to be found in any analytics report, and focusing solely on the big numbers can imply that our efforts have a big impact.
Of course, these “top-of-funnel” metrics don’t really speak to how (or if) your audience actually engages with marketing content, and they can’t tell you if you’re meeting targets for more critical business objectives (like lead generation or ecommerce sales). Even when those KPIs might be in decline, the “top-of-funnel” metrics can help us to paint a rosier picture for ourselves.
But while pageviews and likes might be the most frequently cited examples, there’s really no definitive list out there stating which metrics are and are not “vanity metrics”. And that’s because it’s not the metrics themselves at risk of becoming vainglorious, but us, the people responsible for analyzing and interpreting them.
Any metric can become a vanity metric if we allow it to – even one as trusted, as well-regarded, and as universally relied upon as the “conversion”.
Given the fundamental role that conversions play in pretty much any form of online marketing, I realize that some of you are probably ready to call shenanigans on this whole article right now (if you hadn’t already done as soon as you read the headline).
“Conversion”, after all, is just industry jargon to describe an event whereby “a user completes a desired action”. On its surface, that sounds like a pretty rock-solid measure for the success or failure of our marketing efforts: did the people we’re talking to do the thing we wanted, or not?
However, as with so many areas of online marketing, the devil is in the detail. There are a number of ways that marketers can tend to treat conversions (and conversion rates) more as “vanity” metrics than as “actionable” ones. Here, I’ll be discussing a few of the more common examples that I’ve found in my own experience.
More Attribution ≠ More Conversions
Let’s start with a hypothetical example. Say you’re working on a fairly simple “lead generation” website (the one you’re on right now, for instance). The site includes information about the business’ products and services, a few contact forms, links to their social media profiles, a phone number, maybe a couple of mailto links… you get the idea.
This website already has Google Analytics installed, with goal tracking configured to record any contact form submissions as conversions. In an average month, the site sees… I dunno, let’s say five form-fills. Five conversions per month.
Suppose you decide that you want to start tracking additional goals, above and beyond those form-fills, in Google Analytics. Call tracking is one promising option, in part because it helps to understand how many of a website’s visitors will subsequently place a phonecall to that business.
Callrail is a fairly popular call tracking provider (in fact, it’s the most popular), and their product offers fairly simple integration with Google Analytics. So, you decide to implement call tracking, add a few new “tracking numbers” to the website, and configure a new goal in Analytics which counts any phonecalls placed to these new tracking numbers.
After a full month has gone by, you take a look at your Conversion reports in Google Analytics. It’s been another average month for contact form submissions – five in total – but your new call tracking setup has racked up a total of 25 goal completions, all on its own!
Now, my question is this: would you describe what you’ve just done as “increasing conversions by 500 per cent”? Y’know, like if you were talking to your boss, or updating your LinkedIn profile? Because if you would:
Don’t get me wrong: online marketers should absolutely take steps to improve attribution, there is almost always more work to be done there. And you would be, in the strictest literal sense, correct; that website would have seen five conversions in the month, whereas now it’s seeing 30. But it takes, I think, a certain degree of intellectual dishonesty to claim that you’ve improved the results just because you changed the way you’re measuring them. The comparison serves no one but yourself, and your ego.
I’ve read through enough agency case studies (and audited enough of the Analytics accounts they were based on) to know that this sort of reasoning is all too prevalent among digital marketers. And all it accomplishes, really, is to create the expectation that you’ll be able to repeat that kind of growth rate on an ongoing basis.
You don’t want to set yourself down that road – it ends with your Analytics views set up so that managing to keep a visitor on the website for 15 seconds counts as a “conversion”.
Yes, I’ve seen that too. More than once, and once would have been too often. Just, don’t.
Are Your Lead-Gen Numbers On The Level?
When I’m auditing a new client’s web analytics, one of my main priorities is to understand – specifically – which user actions count towards the “conversions” they’re seeing in their reports.
In the case of ecommerce tracking (when implemented by a competent developer), those actions are pretty cut-and-dried: a conversion gets recorded any time that an online purchase is completed on their website. But when it comes to tracking conversions related to lead generation (contact form submissions, email signups, whitepaper downloads, etc.), things tend to get a little… fuzzier.
Google Analytics offers several different “goal types“, each of which measures conversions based on a different set of criteria. While this built-in flexibility allows the product to be installed and implemented across a wide range of use cases, it can also lend itself to implementations that are less accurate than possible (or desirable) for the marketers who rely on its conversion reporting.
One of my personal pet peeves is when I look into how lead-gen form submissions are being tracked on a business’ website, and find that their Google Analytics “views” measure these conversions based on the number of times that visitors have reached a confirmation (or “thank you”) page.
The problem with using URL-based “destination” goals to track lead generation is that your conversion totals/rates can be inflated by any number of “unexpected” user behaviours. An individual user might refresh the confirmation page, or land on that page multiple times by using the back/forward navigation buttons in their browser. Their Google Analytics session might time-out due to lack of activity (maybe they leave the tab open during their lunch break), and another pageview/conversion gets recorded once they return and a new session begins.
Another common issue: I’ve worked on several websites where some (or all) of these confirmation pages had not been set to noindex. It’s a fairly simple error to make, but what it means is that these “thank you” pages wind up being indexed by search engines, with direct links appearing in search engine results pages (SERPs). Between the human visitors and the bot traffic arriving via those SERP links, the vast majority of your lead-gen “conversions” won’t actually be, y’know, leads generated.
A far more accurate method for tracking lead-gen form submissions (and one which rarely, if ever, requires additional dev work to the website) is to configure a JavaScript “event listener” which waits for confirmation that the form has been validated/submitted, and then triggers a Google Analytics “event” to be fired by the page. A goal can then be set up within your Google Analytics views, based on the number of times this event is sent to and recorded by the property.
Google Tag Manager can help to make the entire setup/testing process for this method fairly straightforward, and GTM’s been available (for free) since 2012. If your lead-gen conversion metrics still rely on destination-based goals, you owe it to yourself to explore event-based tracking instead.
One word of caution: it is common to see your total conversions and conversion rates dip when transitioning from a “destination” to an “event” goal type for tracking form-fills. Reason being, since you’re now tracking actual form submissions, your analytics will be much less prone to the kinds of attribution errors which frequently inflate URL-based goals. This is great news for the accuracy of your marketing data, but perhaps not so great for one’s sense of vanity.
Congrats, You Got A Lead! Now What?
Stop me if you’ve heard this one – a marketing coordinator gets called in for a meeting with the higher-ups, to discuss their performance over the past quarter. They dutifully set about pulling reports, and the numbers all look pretty positive; overall website traffic has been steady, engagement metrics are above the benchmarks, and there’s been an uptick in conversions and conversion rates compared to the same time period last year.
Armed with these stats, the coordinator walks into their meeting with an air of confidence. It’s only once things gets underway that the purpose of their attendance becomes clear: closed deals are down this quarter, revenues are declining, and the Sales team has been complaining about the quality of leads being brought in.
In other words, the marketing analytics are telling us one story, and the balance sheet is telling us another.
However sophisticated your web analytics might be, they can’t tell you everything about the value of your online marketing efforts, or the audiences you’re reaching through them. The reason for this is quite simple: your marketing is not your business. At a certain point, the customers and prospects who’ve interacted with your marketing content will begin connecting with other areas of the business (most frequently, Sales and Support teams), and it’s generally through those interactions that the true value of your audience becomes apparent.
Customer Relationship Management (CRM) systems often serve as the central repository for all of a business’ vital customer data: contact information and details, lead scores, purchase histories, deals closed, records of past interactions with customer service and/or sales teams… the list goes on and on. In order to provide a clear picture of the bottom-line impact your online marketing efforts have, web analytics data should be integrated with the customer data stored in your organization’s CRM database.
If you’re working with Google Analytics (and really, who isn’t?), this integration is typically achieved through the User-ID feature. A User-ID is a unique, persistent identifier assigned to an individual customer record within your CRM, which can then be passed to Google Analytics and associated with a visitor’s session/interaction data stored there. Setting up this integration allows marketers to augment the data being collected/available to them about their website’s visitors, develop a better understanding of the value specific visitors have to their business, and in many cases, it can dramatically improve the accuracy of GA’s user counts.
At this point, you might be asking yourself whether your Google Analytics already has User-ID enabled, or if your marketing team currently takes advantage of the many, many benefits of integrating its web analytics data with CRM data.
You… are not. I’ll say that with 99 per cent confidence: you aren’t doing this right now.
In all my career – working directly with dozens of local businesses of all sizes and industry sectors, auditing hundreds (if not thousands) of websites – I can only think of two examples in Canada where I’ve ever actually seen the Google Analytics User-ID feature implemented and configured. And both of those are banks.
Say what you will about the Canadian banking sector, but they tend to be fairly good at spinning data into profits. Banks make bank.
I’ve never really been sure why there’s such a dearth of Google-Analytics-to-CRM integrations out there in the wild. It might be down to the many online guides which claim that User-ID requires some form of user login to work (though that’s not actually the case). Maybe it’s the fact that Google warns, should you screw up your implementation and start sending them any personally identifiable information (PII) about your site/app’s visitors, that “your Analytics account could be terminated and your data destroyed“.
But more than anything, I think the paucity of these integrations really comes down to marketers being contented with the numbers they already have, and not being terribly fussed about improving their data collection to yield more accurate, more comprehensive insights.
After all, if we were ever to find ourselves in the same position as that marketing coordinator – sat in a meeting with the big-wigs, being asked to justify a slumping ROI – it can be far simpler to just blame Sales for not closing more of the “leads” that our work brings in. Or, if that dog won’t hunt, we might surmise that any problems can be remedied with a “brand refresh”, or a new ad campaign, or by hiring another “22-22-22” to handle the company’s social feeds. In the absence of any credible evidence to the contrary, who’s to argue?
Conclusion: Conversion Confusion Concerns
Fans of Betteridge’s law will be cheered to learn that I don’t actually think of conversions as an a priori “vanity metric”.
What I do think – and I hope I’ve managed to get this across – is that the widespread adoption and acceptance of this metric (combined with the inherent authoritativeness of a term like “convert”) can lend itself to misuse.
The appeal to vanity can be a tremendously effective marketing technique; it works, it’s always worked, and it’s going to keep right on working. That’s probably why we find so many examples of it both within our profession, and in the products of our work. But as digital marketers, it’s worth reminding ourselves from time to time that it’s still a logical fallacy – however compelling and persuasive a fallacy it might be.
Once we begin to treat performance metrics as a means to make our work look good, rather than a means to inform and continuously improve our decision-making, that’s exactly what “conversions” end up being.