Consistency is the part of reporting most agencies are proudest of, and it is also why the video client keeps replying with questions. Somewhere in the effort to standardise, campaign type reporting got flattened into one metric set: impressions, clicks, CTR, spend, maybe CPM, applied in the same order to a brand search campaign, a creator activation on TikTok and a BVOD buy. The layout looks disciplined. What it actually communicates is that you judged all five campaigns by the same yardstick.
You can usually spot this by which reports generate email. The search report has gone out untouched for months because the numbers it shows happen to be the ones that matter for search. The video and influencer reports come back every cycle with “so what does the click-through rate tell us here?”, which is a polite way of asking why you led with a metric that was never the point of the campaign.
The order of a dashboard is an argument about what mattered. Getting it right per channel is mostly a question of deciding what leads, what gets demoted to a table, and what should not appear at all.
The top-left tile is read as a verdict
Clients do not read dashboards left to right, top to bottom, weighing everything equally. They look at the first big number, form an opinion, then scan for confirmation. If the first tile on a YouTube campaign report is CTR at 0.31%, the client has already concluded the campaign underperformed before reaching the completion rate that would have told them otherwise.
That is not a client comprehension problem you can fix with commentary, although good commentary helps. Placement carries more weight than prose. A metric in the hero position is an implicit claim: this is how we think the campaign should be judged. A metric buried in row four of a table is a claim too, that this is context rather than conclusion.
So the question for each campaign type is not which metrics are available (the export decides that) but which one you are willing to be judged on.
Search: report the intent you captured, not the reach you did not control
Search reporting is the easiest of the five because the channel’s own logic does the work. Nobody sees a search ad who was not already looking for something. Impressions are therefore close to meaningless as a performance signal and quite useful as a demand signal: impression volume on non-brand terms tells you how big the category conversation was this month, not how well you performed in it.
Lead with conversions and cost per acquisition, or with cost per click and click-through rate if the account has no conversion tracking worth reporting. Impression share belongs high up when budget is the constraint, because it is the metric that turns “we spent the budget” into “we spent the budget and left 40% of available demand on the table”. That single line justifies more budget increases than any chart of clicks over time.
Demote average position style metrics and raw impressions into a table. Cut anything that implies reach or awareness. A search campaign that reports on reach invites the client to compare it against a display buy it has nothing in common with.
The one search metric worth more space than it usually gets is the query or keyword group breakdown. Clients understand “we captured 61% of clicks on comparison terms, 12% on pricing terms” far better than they understand a blended CPC. The differences between search vs social reporting metrics come down to this: in search you are reporting on demand you intercepted, in social you are reporting on attention you bought.
Social: decide conversion or engagement before you build the row set
Social is where reports go wrong most often, because “social” is two different campaign objectives wearing the same platform name. A Meta Ads conversion campaign and a Meta Ads engagement campaign produce the same export columns and require opposite dashboards.
If the objective was conversions, the report leads with conversions, cost per acquisition and spend, and engagement metrics go into a supporting table where nobody mistakes them for outcomes. If the objective was engagement or reach, lead with reach and frequency, then engagement rate, and resist the urge to include CPA at all. An engagement campaign that reports a €340 cost per acquisition has volunteered a bad number against a target it was never given.
Frequency is the most underused metric in social reporting. It answers the question clients ask in a roundabout way (“are we annoying people?”, usually phrased as “are we still seeing fatigue?”) and it explains falling engagement rates better than creative critique does. When average frequency crosses about 6 or 7 in a short flight, the creative did not get worse, the audience got tired.
Creative-level performance deserves real space here rather than a footnote. Social clients care about which asset worked, and a report that shows the actual creatives next to their numbers ends the separate email asking you to send the winning ad.
Video: completion rates, because nobody clicks a pre-roll
The correct video campaign metrics start from an honest premise: the click was never the goal. Somebody watching a 15 second pre-roll on a phone while waiting for the content they came for is not a person you expected to tap through to a product page.
Lead with view-through rate and completion rate. Then cost per completed view, which is the number that lets a client compare this quarter’s buy against last quarter’s and against the TV budget it was probably taken from. Reach and average frequency follow, because video is usually bought for coverage.
Quartile data (25%, 50%, 75%, 100%) is the part worth presenting as a chart rather than a tile row. The shape tells a story a single VTR number cannot: a campaign that holds 70% of viewers to the halfway mark and then collapses has a creative problem at a specific second, which is actionable. A campaign that loses half its audience in the first quartile has a targeting or placement problem.
CTR can stay in the report, in a table, low down, unremarked. Removing it entirely tends to invite the exact question you were trying to avoid. Showing it in a position that signals “we track this but it is not how we judge the campaign” does the job better. If the client needs the definitions to follow the difference between VTR and completion rate on a skippable format, a shared metrics glossary saves the explanation from being rewritten every month.
Influencer: the metrics a CTR cannot carry
Influencer report metrics are the hardest to standardise because the export is usually thin and the value is partly qualitative. Reach, impressions, engagements, saves, shares and comments arrive from platform screenshots or a creator’s own numbers, and the spend is a flat fee rather than a media cost.
Lead with cost per engagement or cost per thousand reached against the flat fee, because that is the only way a client can compare a €6,000 creator deal against the €6,000 they could have spent on paid social. Then engagement rate against the creator’s own baseline, not against a generic benchmark. A 2.1% engagement rate means nothing in isolation and quite a lot if that creator normally runs at 1.4%.
Saves and shares belong higher than likes. They are the closest thing to intent that this channel produces, and they are what a client’s brand team actually looks at. Comment sentiment deserves a short written note rather than a metric, since counting comments and reading them are different exercises.
Cut CTR unless the campaign used trackable links and the creator actually drove traffic as the objective. Reporting a 0.4% CTR on a brand awareness collaboration tells the client you measured a Reel like a banner.
Display and retargeting: the line nobody asks about
Retargeting almost always looks like the best performing line in the report, and almost always deserves the least credit. It converts people who were already converting. Presenting it as the hero of a mixed-channel report trains the client to want more budget in the one place where incremental value is weakest.
Handle it by demoting it on purpose. Retargeting sits in the channel table with its conversions and CPA visible, honest and unhighlighted, while the prospecting lines get the tiles. If the client asks why retargeting is not featured given how efficient it looks, you now have the conversation you wanted, with the numbers in front of you.
For prospecting display, lead with reach, frequency and viewability if the export includes it, then CPM. Conversions can be shown with a plain note about the attribution window, because display’s conversion numbers are the ones most likely to double count against search.
Campaign type reporting with five metric sets and one template
| Campaign type | Lead with | Demote to a table | Leave out |
|---|---|---|---|
| Search | Conversions, CPA, impression share | Impressions, CPC | Reach, awareness framing |
| Social (conversion) | Conversions, CPA, spend | Engagement rate, reach | Nothing much |
| Social (engagement) | Reach, frequency, engagement rate | Clicks, CPC | CPA |
| Video | VTR, completion rate, CPV | CTR, clicks | Nothing much |
| Influencer | Cost per engagement, CPM on reach | Likes, impressions | CTR without tracked links |
| Display prospecting | Reach, frequency, CPM | Conversions with window noted | CTR as a headline |
| Retargeting | Nothing (keep it in the table) | Conversions, CPA, spend | Hero placement |
The reason most agencies never make this change is the assumption that metrics by campaign type means five templates, five files to maintain and five places for the brand to drift out of sync. That is a fair objection to the version where each channel gets its own document.
It is not the only version available. A dashboard template is a set of rows, and a row is a group-by plus a chosen list of metrics. Build the layout once, then define the search row set, the social row set, the video row set, the influencer row set and the display row set inside it. When you produce a report you pick the right metric group per row rather than picking a different file, and inserting a row set replaces only that row. In Screebl the rates are computed from the totals in your upload rather than averaged across rows, which matters when a video buy has wildly different impression volumes per placement and an averaged VTR would flatter the small ones.
The template locks once reports use it, so a change you make in October for a new client does not quietly reshape the September report that is still sitting in a client’s portal. That is the practical part of the build-once-reuse argument that agencies usually discover the hard way.
Start with whichever campaign type produces the most follow-up email. That is where the current metric order is wrong, and one reordered tile row tends to remove a whole recurring thread. If you want to see the metric grouping in a live template before rebuilding anything, the plans page is the place to start, and the twelve metrics worth showing in any client report is a reasonable baseline for the rows you keep across every channel.
