Vanity Metrics vs. ROI: What Actually Proves Campaign Success

The video above raises the question directly: are the metrics on your report actually the ones that matter to the client, or just the ones that are easiest to make look good. This post works through how to tell the difference.

A vanity metric is not a fake number, it is a real number that does not answer the question the client is actually asking. Impressions, reach, followers, and raw click volume are all real, all trackable, and all capable of looking impressive on a slide while telling a client nothing about whether the money they spent produced a return. The problem is not that these numbers are wrong, it is that they are easy to report on and easy to grow, which makes them tempting to lean on when the metric that actually matters is having a harder month.

A simple test for whether a metric is doing real work

Ask what decision the client could make differently based on this number. Reach tells a client almost nothing they can act on. Cost per qualified lead, or ROAS, tells them whether to spend more, spend less, or shift the mix, which is exactly the decision they are paying an agency to help them make. If a metric does not point toward a decision, it belongs lower in the report, not at the top of it.

• Lead with metrics tied to revenue or qualified pipeline, not just activity or exposure
• Keep vanity metrics in the report for context, but do not let them anchor the narrative
• Explain the connection between the outcome metric and the client's actual business goal explicitly
• Be as clear about a metric that moved the wrong direction as one that moved the right way

Why this discipline protects the relationship long term

A client who has been shown vanity metrics as the headline for two years eventually starts asking harder questions, usually right when the agency can least afford it, at renewal or during a budget review. A client who has been shown accurate outcome metrics all along, including the months that were not great, trusts the report enough that a single soft quarter does not put the whole relationship at risk. This is a large part of why reporting built around outcome metrics rather than activity metrics tends to hold accounts through the periods when the numbers are not the story an agency wants to tell.

The usual suspects, and why each one is easier to grow than to trust

• Impressions and reach, which grow with spend almost by definition and say nothing about whether the audience reached was the right one
• Follower and audience counts, which correlate weakly with revenue and can be inflated by contests, giveaways, or paid follower campaigns that attract the wrong people
• Engagement rate, which rewards content that is easy to react to over content that actually moves someone toward a purchase decision
• Session duration and pages per visit, which can just as easily indicate a confusing site as an engaged one
• Branded search volume, which often reflects demand the business would have captured anyway rather than demand the campaign created

None of these numbers are worthless. They are diagnostic: a sudden drop in engagement rate can flag a creative fatigue problem worth investigating, and a jump in branded search after a campaign launch is a reasonable signal that awareness moved. The mistake is not tracking them, it is presenting them as the proof of value instead of the supporting detail, which is a promotion they have not earned.

When ROAS itself is the vanity metric

This is the part that surprises agencies that already think they have solved the vanity metrics problem by leading with ROAS instead of reach. Platform-reported ROAS is frequently inflated, view-through conversions get counted alongside click conversions, attribution windows get set generously by default, and a platform grading its own homework has every incentive to report a number that makes continued spend look justified. A client comparing platform-reported ROAS across two channels is often not comparing apples to apples at all, because each platform sets its own rules for what counts as a conversion and when.

The more rigorous question is not what did the platform report, it is what would have happened without the spend, which is the concept behind incrementality. A campaign that reports strong ROAS but is mostly capturing demand that already existed, branded search, direct visitors who were coming anyway, is not creating the value the number implies. This does not mean every account needs a full incrementality study, most do not have the volume to justify one, but it does mean treating a single-platform ROAS number with some skepticism, and triangulating it against the client's own sales data whenever that data is available.

Moving a client off a vanity-metric habit without a confrontation

A client who has been shown reach and engagement as the headline for years did not arrive at that habit on their own, an agency taught them to look at those numbers, often unintentionally, by leading a report with them for a long time. Abruptly stripping those metrics out of the next report and replacing them with unfamiliar outcome metrics tends to read as evasive, even when the intent is the opposite, because the client notices the numbers they know how to read have disappeared.

The more durable fix is additive, not subtractive. Introduce the outcome metric alongside the familiar one for a full reporting cycle or two, explicitly draw the connection between them, then let the report's emphasis shift naturally as the client starts asking about the new number unprompted instead of the old one. By the time the vanity metric quietly moves to a smaller line in an appendix, the client has already made the mental shift themselves, which sticks far better than being told which numbers to care about.

A simple three-tier structure for the report itself

• Primary: the one or two metrics tied directly to revenue or qualified pipeline, given the most visual weight on the page
• Secondary: leading indicators that reliably predict the primary metric a few weeks out, useful for showing momentum before the outcome catches up
• Context: the reach, impression, and engagement numbers, present for anyone who wants them but never the headline

This structure does most of the persuading on its own, because it makes the hierarchy visible rather than leaving the client to guess which number matters most from how big the font is. An agency that can point to a report and say this is what we lead with and here is why, every single month, builds a kind of credibility that a single strong quarter of vanity metrics never will.

A scenario that shows the gap in practice

A paid social account can illustrate the problem cleanly. Engagement rate climbs for a few months, comments, shares, and saves all trending the right direction, and the monthly report looks strong on the surface. Meanwhile the sales team is quietly reporting that leads from the channel have gotten harder to close, more tire-kickers, fewer people who actually fit the client's ideal customer profile. If the report only tracks engagement, that tension never surfaces until the client notices it independently, usually at the worst possible time, and starts asking why a channel that looks so healthy on paper is not producing revenue.

The fix is not to abandon the channel or the creative approach that is driving engagement, it is to pair the engagement number with a lead-quality metric from day one, so the two can be read together. Rising engagement paired with falling lead quality is a specific, actionable signal, it usually means the creative is optimized for reactions rather than for the audience segment that actually converts, and it points toward a concrete fix: tightening targeting or adjusting the call to action, not just declaring the channel is not working.

Catching that pattern early, inside a single reporting cycle rather than a full quarter, is the entire argument for tracking both numbers side by side from the start rather than adding lead quality only once a client raises a complaint. A report built to surface the tension automatically is worth more than an agency that has to be prompted to go looking for it.

Metrics that deserve a second look before getting labeled vanity

Not every reach or engagement number is automatically decorative, and treating them that way in every context is its own kind of oversimplification. For a media or content-driven business whose revenue model runs on advertising or sponsorship, audience size and engagement are closer to the primary metric than the vanity one, because the audience itself is the product being sold to advertisers. For a brand-awareness phase of a campaign, ahead of a product launch with no sales mechanism live yet, reach and recall are legitimately the leading indicators worth reporting, as long as the report says so explicitly rather than implying they are proxies for revenue they are not yet connected to.

The discipline is not banning a category of metric outright, it is being precise about what each number is standing in for and saying so plainly in the report, so the client never has to guess whether a number is the point or just the supporting evidence. Get that labeling right once, at the start of the account, and it rarely needs revisiting, because the client learns to ask which tier a new metric belongs in before assuming it belongs at the top of the page.