Skip to content

Digital Marketing Foundations · KPIs and unit economics · lesson 10 of 15 · 14 min

Reporting and dashboards: a weekly view that drives decisions

Reports exist to change decisions

A report that nobody acts on is decoration. A good marketing report answers three questions in under two minutes: What happened? Why? What will we do next? Everything else is supporting detail for the people who want it.

The three layers of a useful dashboard

| Layer | Audience | Metrics | Rhythm | |---|---|---|---| | Business | Owner, leadership, client | Revenue, new customers, nCAC, MER, contribution after marketing | Weekly and monthly | | Channel | Marketer or agency | Spend, results, CPA or ROAS by channel, email and WhatsApp revenue | Weekly | | Diagnostic | The person running campaigns | CPM, CTR, CVR, frequency, landing page view rate, hook rate | Daily glance, weekly review |

Start with the business layer. If the top line is healthy, the diagnostics are for curiosity; if it is not, they tell you where to look.

Hands-on: a weekly dashboard in Google Sheets

Create a tab called Data where each row is one week and one channel:

| Week | Channel | Spend | Revenue | Orders | New customers | |---|---|---|---|---|---| | 2026-09-01 | Meta | 1200 | 3900 | 52 | 41 | | 2026-09-01 | TikTok | 800 | 1900 | 25 | 22 | | 2026-09-01 | Email | 60 | 2100 | 30 | 4 |

Then a Summary tab with the week in cell B1:

Total spend           =SUMIFS(Data!C:C, Data!A:A, $B$1)
Total revenue         =SUMIFS(Data!D:D, Data!A:A, $B$1)
New customers         =SUMIFS(Data!F:F, Data!A:A, $B$1)
MER                   =IFERROR(B3/B2, "")
nCAC                  =IFERROR(B2/B4, "")
Meta CPA              =IFERROR(SUMIFS(Data!C:C,Data!A:A,$B$1,Data!B:B,"Meta")
                        / SUMIFS(Data!E:E,Data!A:A,$B$1,Data!B:B,"Meta"), "")
Change vs last week   =IFERROR(B3/SUMIFS(Data!D:D, Data!A:A, $B$1-7) - 1, "")

Revenue should come from your store or CRM, not from ad platforms, so MER and nCAC are platform-neutral. Platform-reported conversions can sit in separate columns for within-platform optimisation.

When you outgrow a spreadsheet, Looker Studio (free) connects to Google Sheets, GA4, Google Ads and Search Console and turns the same logic into charts with a date picker. Connectors for Meta and TikTok usually require a third-party connector or a data export.

Visual rules that make dashboards readable

  • Numbers first, charts second: a scorecard row (spend, revenue, new customers, MER, nCAC) with week-on-week change.
  • One chart per question: a line for trends over time, a bar for comparing channels, a table for detail. Avoid pie charts with more than a few slices.
  • Always show a comparison: last week, the previous four-week average, or target. A number without context is meaningless.
  • Label definitions: "Revenue = store revenue incl. VAT, excl. refunds".

The written summary: what, why, next

Under the dashboard, write three short bullets:

  • What: "Revenue £9,400 (+12% vs 4-week average); nCAC £31 (target £35)."
  • Why: "New TikTok creator video drove 40% of new customers; Meta CPM rose 18% ahead of Black Friday."
  • Next: "Brief two more creator videos in the same style; cap Meta prospecting at current spend until CPM settles; fix the slow checkout page on Android."

Using AI to draft the summary (and not the facts)

AI assistants are useful for turning a table into a first-draft commentary:

Here is this week's marketing table (CSV) and the previous four weeks.
Write a summary in three bullets: What happened, Why (only reasons supported
by the data or my notes below), Next actions. Use plain English for a busy
business owner. If a change could be random noise, say so. Do not invent causes.
Notes: new TikTok creator video launched Tuesday; checkout page slow on Android.
[paste CSV]

Check every number it quotes against the sheet, and remove any "reason" it could not know. Do not paste customer-level personal data – aggregated weekly totals are enough.

Worked example: an agency's client report

A Dubai agency used to send clients a 25-page PDF of platform screenshots. Clients skimmed it and asked, "So are we doing well?" The agency switched to a one-page Looker Studio dashboard (scorecards, a 12-week MER trend, spend and nCAC by channel) plus the three-bullet summary, and a 15-minute weekly call. Decisions now happen in the call, and the agency spends less time on reporting and more on testing.

Decision rules belong next to the dashboard

Write rules in advance so the dashboard triggers action, not debate:

  • If nCAC exceeds target for two consecutive weeks, review the weakest channel's funnel metrics before changing budgets.
  • If MER falls while platform ROAS rises, check attribution overlap and run a holdout test.
  • If a channel beats target for three weeks, increase its budget by 20–30% and watch nCAC.

Common mistakes

  • Reporting platform screenshots instead of business results.
  • No comparisons, targets or definitions.
  • Mixing time zones or currencies between sources.
  • Letting AI write "insights" that the data does not support.

How to measure success

A dashboard works when decisions happen faster and are written down: count the actions agreed each week, and check whether they were done by the next report.

Video lecture: Reporting that drives decisions: what, why, next

Lecture coming soon · 11 chapters · about 8 minutes. Read the full transcript below.

  1. Reporting that drives decisions
  2. Reports change decisions
  3. Three layers
  4. Hands-on: the Data tab
  5. Always compare
  6. Visual rules
  7. Example 1: three-bullet summary (illustrative)
  8. AI-drafted summaries
  9. Example 2: Dubai agency (illustrative)
  10. Decision rules
  11. Recap and try this now

Lecture transcript

Reporting that drives decisions

Here's a confession from a lot of agencies. They used to send clients a twenty-five page PDF every month, full of screenshots and colourful charts. And every month, the client replied with one question: so, are we doing well? That's the sign of a report that doesn't work. In this lecture you'll learn how to build a weekly marketing dashboard that answers three questions in under two minutes: what happened, why, and what we'll do next. You'll build it in Google Sheets with a handful of formulas, learn the visual rules that make it readable, and see how to use AI to draft the commentary without letting it invent facts.

Reports change decisions

Here's the key idea. A report exists to change decisions. If nobody does anything differently after reading it, it's decoration. So before you add a single chart, ask: what decision will this number inform? Think of a car dashboard. It doesn't show you the temperature of every bolt in the engine. It shows speed, fuel, and a warning light when something's wrong. Your marketing dashboard should work the same way: a few numbers you check every time, and warning lights that tell you where to look. The engine diagnostics are there when a warning light comes on, not before.

Three layers

A useful dashboard has three layers. The business layer is for owners, leaders and clients: revenue, new customers, new-customer CAC, MER, and profit after marketing. The channel layer is for marketers: spend, results and cost per result or ROAS for each channel, including email and WhatsApp. And the diagnostic layer is for the person running campaigns: CPM, CTR, conversion rate, frequency and hook rate. Always start at the top. If the business layer is healthy, diagnostics are for curiosity. If it isn't, they tell you where to dig.

Hands-on: the Data tab

Now let's build it. In Google Sheets, make a tab called Data where each row is one week and one channel: week, channel, spend, revenue, orders and new customers. Here's the important part. Revenue and new customers should come from your store or CRM, not from ad platforms, because platforms each credit themselves and their totals overlap. Then make a Summary tab. Put the week you're reporting in one cell, and use SUMIFS to total spend, revenue and new customers for that week. MER is revenue divided by spend. New-customer CAC is spend divided by new customers. All the formulas are in the lesson text, ready to copy.

Always compare

Let's add one more formula that makes the dashboard come alive: change versus last week. Divide this week's revenue by last week's, subtract one, and format it as a percentage. Do the same for new customers and CAC. Now you have context, and context is everything. A number on its own is meaningless. Is eight thousand pounds good? Compared with what? So always show a comparison: last week, the previous four-week average, or the target. When your business grows, Looker Studio, which is free, can connect to your sheet, GA4, Google Ads and Search Console, and turn the same logic into charts with a date picker.

Visual rules

A few visual rules will make your dashboard readable. Numbers first, charts second: a row of scorecards at the top. One chart per question: a line for trends over time, bars to compare channels, and a table for detail. Avoid pie charts with lots of slices, because people can't compare angles. Label your definitions, like revenue equals store revenue including VAT, excluding refunds. And keep time zones and currencies consistent across sources, or your numbers will never match. Simple rules, but they separate dashboards people use from dashboards people ignore.

Example 1: three-bullet summary (illustrative)

Now the written summary. Under the dashboard, write three bullets. Here's a simple worked example for an online store, with illustrative numbers. What: revenue nine thousand four hundred pounds, up twelve percent on the four-week average, and new-customer CAC thirty-one pounds against a target of thirty-five. Why: a new TikTok creator video drove about forty percent of new customers, and Meta CPMs rose eighteen percent ahead of Black Friday. Next: brief two more creator videos in the same style, hold Meta prospecting spend steady until costs settle, and fix the slow checkout page on Android. That's the whole report. A busy owner reads it in thirty seconds.

AI-drafted summaries

Can AI help? Yes, with the drafting, not the facts. Paste this week's aggregated table and the previous four weeks into an assistant, add your own notes about what changed, and ask for three bullets: what happened, why, and next actions. Tell it to use only reasons supported by the data or your notes, and to flag changes that could be random noise. The full prompt is in the lesson text. Then check every number it quotes against your sheet and delete any reason it couldn't possibly know. And don't paste customer-level personal data. Weekly totals are all it needs.

Example 2: Dubai agency (illustrative)

Now a realistic business scenario. A Dubai agency replaced its twenty-five page monthly PDF with a one-page Looker Studio dashboard: five scorecards, a twelve-week MER trend, and spend and new-customer CAC by channel, plus the three-bullet summary. They also added a fifteen-minute weekly call. The result? Clients stopped asking, are we doing well? because the first line answered it. Decisions now happen in the call and get written down. The agency spends less time building reports and more time running tests, and clients stay longer because they can see the link between the work and the results.

Decision rules

Put decision rules next to the dashboard, so numbers trigger action instead of debate. For example: if new-customer CAC is above target for two weeks in a row, review the weakest channel's funnel metrics before touching budgets. If MER falls while platform ROAS rises, check attribution overlap and consider a holdout test. If a channel beats target for three weeks, raise its budget by twenty to thirty percent and watch CAC. And avoid the classic mistakes: platform screenshots instead of business results, no comparisons or definitions, mixed time zones and currencies, and AI insights the data doesn't support.

Recap and try this now

To recap. Reports exist to change decisions. Build in layers: business first, then channel, then diagnostics. Use store or CRM revenue for MER and new-customer CAC, always show a comparison, and label your definitions. Write a three-bullet summary: what, why, next. Let AI draft it, but you own every number and every reason. How will you know it's working? Decisions happen faster and they get done by the next report. Here's your try this now. Build the Data and Summary tabs from the lesson text with four weeks of data, real or illustrative, and write your first what, why, next summary.

Key takeaways

  • A report must answer what happened, why and what we will do next.
  • Layer dashboards: business metrics first (revenue, new customers, nCAC, MER), then channel, then diagnostics.
  • Use store or CRM revenue for blended metrics, always show comparisons and label definitions.
  • AI can draft the summary from aggregated data, but you check every number and remove unsupported reasons.

Try it

Build the Data and Summary tabs in Google Sheets for four weeks of real or illustrative data, then write the three-bullet what/why/next summary for the latest week.