Marketing Vanity Metrics and Which Numbers Predict Revenue

The report lands on the first Tuesday of the month, click-through rate up again, cost per click down for the fourth period running. The owner reads it twice, then opens the bank account, which looks roughly the way it looked in month one. Nothing in the report is false. It just never reaches the part of the business that pays salaries. That gap is what people mean by marketing vanity metrics, and the argument about them usually gets framed as honesty against dishonesty when the real question is depth: how far down the chain from impression to collected cash does a number sit, and what actually moves when it moves.

Here is the diagnostic we run on a reporting pack before anyone argues about media. Take each number on the page, imagine it improving by half again overnight, and write down what that improvement forces to happen next. Some numbers dead-end inside two steps. Others drag revenue along behind them whether anyone is watching or not.

The marketing vanity metrics test is one question

Cheaper clicks buy more clicks. More clicks only matter if the rate at which they become real conversations holds while volume rises, and it frequently does not, because the cheapest inventory tends to be the least interested. LocaliQ's 2026 search benchmark study quotes a performance marketing SVP making the point without any moralising: CPC and CTR "are health metrics", worth keeping a pulse on and worth using as levers, "However, they aren't KPIs", and improving them at the expense of return on advertising spend is a bad trade (LocaliQ).

That is the distinction worth keeping. Health metrics belong in the report, on the diagnostic page, where they tell a buyer which part of the account is sick. They do not belong on the first slide an owner reads for two minutes before a leadership meeting.

The same study puts the 2026 cross-industry average search CTR at 6.64% and the average Google Ads conversion rate at 8.18% (LocaliQ). Those are worth knowing for shape and worthless as a pass mark. They average a wide set of mostly consumer industries with no separate cut for high-ticket considered purchases, so a firm selling a $25,000 engagement cannot read its own health off them in either direction.

Confidence in measurement is not measurement

Ask an agency whether they track ROI and the answer is always yes. Nielsen puts a number on how little that answer is worth: 85% of marketers say they are confident in their ability to measure ROI, while only 32% actually measure it holistically across traditional and digital media channels (Nielsen). The same research finds 38% of marketers now name sales or ROI as their top metric for success (Nielsen), which is movement in the right direction and still leaves plenty of reports anchored somewhere shallower.

So replace the yes/no question with two mechanical ones. Which system holds the closed-won number, and who types it in. Then: does that number ever travel back to the ad platform, or does it sit in a CRM the media team never opens. An agency that cannot answer in a sentence is reporting on the half of the funnel it can see, which is a competence gap rather than a character flaw. Settling who owns that answer up front is part of the difference between a growth partner and a vendor.

The platforms already sell the revenue version

Google's own documentation draws the line. Value-based bidding optimises campaigns on the value brought to the business, maximising conversion value within a given budget or using target ROAS, while conversion-based bidding maximises conversion volume within a budget or against a target CPA (Google Ads Help). To run the first, you have to report two or more different values to your conversion goals, either real economic values like revenue or proxy values like a lead score (Google Ads Help).

That requirement is where most accounts stop. Scoring leads or passing back deal values means someone has to connect the CRM, agree what each lead type is worth, and keep the mapping current as the offer changes. It is a week of unglamorous work, and skipping it leaves the account optimising toward whichever form gets filled in most often.

Offline conversion imports cover the other half. Google's documentation describes the common case where an ad starts a customer down a path that ends in a sale at your office or over the phone rather than online, and issues a unique click ID (GCLID) for every click that reaches your site from an ad so you can store it with the lead and send the outcome back later (Google Ads Help). Google also reports that advertisers who used first-party data such as email addresses and phone numbers alongside GCLIDs in offline import saw a median 10% increase in conversions compared with those using standard offline conversion imports (Google Ads Help). Read that one carefully. It describes more of the sales you already made getting counted, self-reported by the platform about its own product, and says nothing about incremental revenue. An agency quoting it as a revenue lift is selling.

A lead count flatters deals with many stakeholders

Lead volume survives longest of all the shallow numbers because it feels close to revenue. It is a count of individuals, though, and Forrester's Buyers' Journey Survey, 2024 found that on average 13 people are involved in making a purchasing decision (Forrester). Where the buyer is a founder with a card, the individual and the decision are the same event. Where you sell into a committee, a rising lead count can describe the same handful of accounts sending more people to your site.

Forrester's argument is to pivot from individual leads toward engaging multiple members of the buying group. The post carries one customer's self-reported results from doing it: win rates doubled during the pilot stage, and an 800% increase in opportunity progression to forecast once it scaled (Forrester). That is one account describing its own programme. Take it as an illustration of what changes when the unit of measurement changes, and not as a number you should expect to reproduce. The practical version for a smaller operation is simply to count accounts and conversations rather than email addresses, which is also a useful lens on the constraint described in this piece on lead-gen campaigns that stop scaling.

Build the scorecard, then give it a lagged window

The pressure to prove value pushes people toward the shallow end by itself. Duke's CMO Survey, which polled 308 marketing leaders at for-profit U.S. companies, 97% of whom are VP-level or higher, found more than 70% reporting that they prioritise immediate results over long-term gains, often relying on established strategies rather than new investments (The CMO Survey). The same survey notes that acquisition budgets remain 26% larger than retention budgets despite retention delivering stronger performance outcomes than acquisition (The CMO Survey). A weekly revenue review recreates that pressure inside your own building.

Four lines, reviewed monthly: qualified conversations held, opportunities created, closed-won revenue credited to the month the lead arrived rather than the month it closed, and cost per acquisition set against contribution margin. Underneath them, keep the two leading indicators that genuinely move those four, speed to lead and show rate on booked calls. CPC, CTR, cost per lead and impressions stay in the pack, on the diagnostic page, where the media buyer uses them for what they are.

Run this on your own numbers

Assume $30,000 of monthly spend at a $12 average CPC, which buys 2,500 clicks. Assume 6% of those clicks become a form fill, so 150 leads. Then pick your own funnel rates and substitute them; this model runs on a 30% booking rate (45 bookings), a 60% show rate (27 calls held), a 20% close rate on held calls (5.4 deals), and an $18,000 average contract value. That is $97,200 of new revenue against $30,000 of spend.

Now pull the vanity lever. The buyer gets CPC down 20% to $9.60, which buys 3,125 clicks, and the report looks excellent. If the cheaper traffic converts at 4% rather than 6%, that is 125 leads, fewer than you started with, and the same downstream rates produce 4.5 deals worth $81,000.

Pull a revenue lever instead. Leave the media alone and lift show rate on booked calls from 60% to 75%. The same 45 bookings become 33.75 held calls, 6.75 closes and $121,500. The number that moved there is owned by the calendar and the follow-up sequence, which is the whole reason it earns a line on the scorecard while CPC does not. Run the same three passes on your real rates before your next agency review.

Hold your agency to the numbers that survive the doubling test, and hold yourself to giving those numbers a window long enough to show up in. One month of closed-won revenue by lead cohort tells you more than a quarter of click-through charts, and it will not say anything honest until the sales cycle plus a couple of weeks has passed. When the pack cannot show where the money landed, the fix is usually the plumbing between the ad account and the CRM rather than a new logo on the reporting deck. If you want a second read on which lines in your current pack are load-bearing, that is the first thing we do on a call.