Marketing Dashboard

Marketing numbers do not live in one place. Ad spend sits in the ad account, visits in a web analytics tool, signups in the product database, revenue in the payment processor. Each screen has its own date picker and its own idea of how to count a week.

So the night before the weekly meeting always looked the same. Somebody opened five tabs, copied numbers into a sheet, built a table, and compared it to last week. That took about ninety minutes, and a different person did it slightly differently every time.

"Last week's deck said 420 signups. This one says 390." Half the meeting goes to explaining the gap. One person counted only signups attributed to ads and the other counted every signup, and neither table said which. Nobody was wrong, and nobody could prove it inside the meeting, so the decision that needed making got pushed to next week.

A marketing dashboard removes the copying and nails the definitions to the screen. Gathering the numbers in one place is half of it. The other half is deciding once what to look at and under what definition, then leaving it alone. Seeing the same metrics in the same order every week is the point of the thing, not a side effect.

It removes three problems that come with hand-built tables.

1. Copying — connect each tool once and the numbers arrive on their own. The ninety minutes before the meeting disappear, and so does the transposed digit that nobody catches until someone asks about it in the room.
2. Drifting definitions — what counts as a signup, and whether the week starts Monday or Sunday, gets fixed on the screen. A new hire two quarters from now reads the same number the same way.
3. No basis for comparison — last week, last month, and the same month last year appear beside today. A single figure of 400 tells you nothing; four weeks side by side settle the question in a second.

The metrics on it usually follow the funnel: acquisition, conversion, cost, and return.

1. Acquisition — visits by channel. The raw material for splitting by channel is the UTM parameter, a tag appended to a link that records which route someone arrived through. Campaigns sent without one collapse into an unattributed pile in this row.
2. Conversion — the conversion rate, the share of visitors who take the target action. Four thousand visits and 180 signups is 4.5 percent.
3. Cost — customer acquisition cost (CAC), what one customer costs to win, alongside ROAS, revenue earned per unit of ad spend.
4. Return — what the customers who arrived actually end up paying. This row differs from the other three in that it never settles inside the week, which is why a cohort analysis view usually sits next to it, grouping people by when they arrived.

The hard part is not building the screen but cutting it down. Twenty metrics and nobody looks; five and the row that moved is obvious thirty seconds into the meeting. To decide what to cut, ask who does what differently when a given number moves. A row with no answer is a vanity metric — impressions being the usual example, since nothing follows from them going up or down. Cutting is unpopular because every row had a reason to be added once, so it helps to review the screen on a schedule, say once a quarter, rather than arguing about it whenever someone proposes an addition.

Three other ways cover the same need, and which one fits depends mostly on team size.

1. A spreadsheet — copy the numbers in by hand each week. With two channels and five metrics this is faster, and in the early months, when what you want to watch changes monthly, a sheet bends more easily than a wired-up screen.
2. Reading the tool screens directly — open the ad account and the analytics tool as they are. Fine while one person owns the numbers; the moment three people join the meeting, all three are looking at different tabs.
3. A recurring report — a document that summarizes the state of things on a schedule. It beats a screen in that a human can attach interpretation to the numbers, and loses to a screen in that a human has to write it every time. In practice teams use both: read the dashboard, write only the interpretation in the report.

If a company-wide dashboard already exists, the marketing view often becomes a tab inside it rather than a separate product, especially once the underlying data lands in one warehouse.

"That number's on the dashboard — have a look before the meeting."

A request to stop rebuilding the same table in three different decks.

"Before we add metrics to the dashboard, let's drop the ones nobody reads."

A proposal to keep the screen short enough that people actually read it.

"CAC looks off on the dashboard this week — ad spend lands a day late."

The anomaly is a data timing issue, not a change in performance.

At Pagedeck, Ravi and Mia started by cutting the screen from eight rows to five. They used one test for what stayed: if this row moves, can we say who changes what this week? Impressions and follower count had no answer and came off. Signups by channel and cost per acquired customer had one and stayed. At the bottom of the screen they wrote out the definition and the counting window for each remaining row, so that whoever joins in six months doesn't quietly read the same numbers a different way.

The most common misread is that having a dashboard means performance is being managed. Visible numbers feel like control. But the screen reports a state; why the state changed lives one level down, in the attribution rules or in the individual campaign records. Attribution is the rule that decides which touchpoint gets credit for a conversion. Teams that keep the screen open but never open what sits under it end a meeting having confirmed that signups fell, and nothing else. The screen is worth having for exactly that reason: it tells you which question to go ask.

The second misread is that more numbers make a better screen. Loading every metric the tools offer looks thorough, but a reader facing twenty rows cannot tell which one matters today. Decisions quietly go back to instinct, and within a few weeks nobody bothers putting the screen up at all.