Marketing Reports & Dashboards

A report is an instrument for making a decision

Marketing reports exist to move a decision from opinion to evidence. That purpose sets the format, the frequency and the contents, and it is why a single marketing dashboard rarely works: different decisions need different numbers at different intervals.

Separate two kinds of metric first, because most packs mix them without labelling. An activity metric describes what the team did: emails sent, posts published, campaigns launched, impressions bought. An outcome metric describes what changed in the market: qualified conversations, opportunities opened, pipeline value, revenue, win rate. Activity metrics are useful for diagnosis and for catching execution failures. They are not evidence of effect, and presenting them as evidence is how marketing loses credibility with a finance-minded audience.

Distinguish also a report from a dashboard. A report is a narrative delivered on a schedule by a named author who has interpreted the numbers. A dashboard is a self-serve surface for people who already know what they are looking for. Most organisations build the dashboard first, find nobody opens it, and conclude they have a data problem.

Four audiences, four documents

Write for one reader at a time. The four common audiences want incompatible things.

• The board. Quarterly, one page, trend over time, market position, and a small number of outcome metrics tied to the plan. Directors want to know whether the strategy is working and whether the money bought anything. Channel detail is noise here.
• The chief executive. Monthly: pipeline contribution against plan, cost of acquiring a customer and how long it takes to pay back, and any material change in demand.
• The sales leader. Weekly, organised by account rather than by channel. Which target accounts showed activity, which enquiries were worked and how fast, which sources produce opportunities that survive qualification. If it cannot be read as a list of accounts, sales will not read it.
• The channel manager. Weekly, granular, diagnostic, and full of the activity metrics the board should never see: creative fatigue, bid changes, deliverability, landing-page behaviour.

Four documents drawn from one dataset is maintainable. One document trying to satisfy all four is a compromise everybody distrusts.

Leading indicators, lagging indicators, and being judged fairly

In a long cycle, outcome reporting alone is a trap for the marketing team. If revenue is the only accepted measure and revenue arrives two or three quarters after the work, the team is permanently assessed on a period it can no longer influence, and the budget conversation happens with the wrong data on the table.

Report both, explicitly labelled. Lagging indicators are the verdict: closed revenue, win rate, average contract value, payback period. Leading indicators are the evidence available now that the verdict will be favourable later: qualified conversations created, coverage of the target account list, named accounts showing repeat engagement, opportunities created and their stage progression, and the share of new pipeline sourced rather than merely passed through.

Two rules keep this honest. A leading indicator must be something you can be wrong about; a number that only ever rises is decoration. And every leading indicator needs a stated hypothesis linking it to a lagging one, so that when pipeline fails to follow engagement the assumption is visible and can be revised rather than quietly dropped.

The metrics that flatter, and why they survive

Vanity metrics persist because each has a real reason for existing. Naming the reason beats moralising.

• Impressions. Large, cheap to increase, and present in every platform by default. An impression is an opportunity to be seen by someone who may not be in your market. Report reach against the target account list, or cost per qualified conversation, instead.
• Sessions. They sit on the front page of every analytics tool and correlate loosely with everything. In a market of a few thousand buyers, total sessions largely measures competitors, students and bots. Segment to the pages and accounts that matter.
• Followers. The number is cumulative and never falls, which makes it a comfortable slide. It measures accumulated history, not current performance.
• Open rate. It is the only number available between sending and clicking, and it is inferred from a tracking image loading; where a mail client pre-fetches or blocks that image the figure stops describing human behaviour. Treat it as a rough deliverability signal and judge campaigns on replies and meetings.

Retiring a metric is a political act. Do it by replacing it in the same slot, in the same report, with the number you would rather be judged on.

Cadence, noise and manufactured urgency

Match the reporting interval to the rate at which a metric can carry signal. Weekly reporting on a metric whose month-to-month variation is mostly random manufactures urgency, invites intervention, and produces the worst failure mode in marketing management: changing the campaign every fortnight and never learning whether any version worked.

Small numbers are the reason. A page that converts three visitors out of eighty, then six out of ninety, has not doubled its performance; that is what small counts do. Report the count alongside the rate, always, and where the denominator is small say plainly that the range of plausible values is wide. A board pack does not need a formal confidence interval. It needs a sentence: this is based on 14 enquiries, so treat a change of two or three as noise.

A workable default is operational metrics weekly, outcome metrics monthly on rolling twelve-month windows, and strategic assessment quarterly. Then hold the line. The most valuable discipline in reporting is refusing to interpret a number that cannot yet be interpreted, and explaining why.

Annotate, or the number has no cause

A number without a cause invites invention, and the invention is usually flattering. Keep a dated change log beside the report and mark the events on the charts. What belongs in it:

• Campaign starts, stops and budget changes, expressed as proportions rather than buried figures.
• Website releases, template changes and navigation edits.
• Tracking changes: a tag broken or fixed, a consent banner deployed or reconfigured, a retention setting changed, a conversion definition edited. These cause the largest unexplained steps in analytics data and the most confident wrong explanations.
• Pricing, packaging and positioning changes.
• Sales-side events: headcount and territory changes, a hiring gap, one unusually large deal.
• External events: a competitor's funding announcement, a regulatory deadline in your buyers' industry, a trade show.

Annotation is also the honest defence against being blamed for a step you did not cause. A tracking migration that halves recorded conversions looks identical to collapsing demand if nobody wrote down the date.

The one page, and the dashboard nobody opens

The one-page monthly marketing KPI report is the most reliable format available. On it: three to five outcome metrics against plan, a matching set of leading indicators, one chart on a rolling twelve-month window, the annotations for the period, a paragraph on what changed and why, and a paragraph naming the decision you want the reader to make or endorse. Nothing else. Anything needing an appendix goes in a separate document nobody is obliged to read.

Automate the data collection and keep the interpretation human. Scheduled exports, a warehouse and a template repay their setup cost; automated commentary does not. A dashboard refreshing hourly and opened twice a year is worse than a monthly slide someone reads, because it creates the impression that reporting is handled.

Two closing practices. Give every recurring report a named owner who presents it and answers for it. And review the report itself annually: strike any metric that has never changed a decision, and ask each audience what question they wanted answered that the report did not address.

Frequently Asked Questions