Anomaly AI LogoAnomaly AI

Marketing Analytics — Guide · 2026

Marketing Dashboard: Examples, KPIs and How to Build One

Pick the KPIs your audience needs, adapt one of six example layouts, build it from your data and check the numbers before you share it.

Ash RaiUpdated 13 min read

Quick answer — what a marketing dashboard should include

A marketing dashboard is one screen for one audience, showing the few numbers that audience uses to judge marketing: spend, results and efficiency by channel, each against a target. Write down each KPI's formula and source before you build. Before you share it, recompute ratios from totals and drop incomplete days. On our downloadable 90-day sample, averaging a row-level ROAS column turned a true 2.88 into 4.03.

The monthly review is on Thursday. Your director wants one screen: are we on target, and which channel is carrying the month? You have exports from the ad platforms, the email tool and web analytics, and last month's spreadsheet averaged a ROAS column that nobody can vouch for.

Below you'll find what to include, six example layouts and a build you can copy. Every number can be checked in a downloadable 90-day sample.

Who this is for: this guide is for marketers and consultants who need to show a manager or a client how each channel performed, in one place. Their data sits in spreadsheets, GA4, or ad accounts such as Google Ads, Meta and TikTok.

What is a marketing dashboard?

A marketing dashboard is a single view of the numbers a team uses to judge its marketing. It shows what was spent, what came back and how efficiently, usually by channel and against a target. A marketing analytics dashboard or digital marketing dashboard is the same idea, built from web analytics, ad platforms, email and search data. A marketing reporting dashboard is the version you hand to someone else to read.

Dashboards in marketing are used for three jobs. Teams check progress against a target, find the channel or campaign that needs attention, and give managers or clients a view they can open themselves. Deciding between a dashboard and a report? Read dashboard or report: which to send stakeholders.

Which KPIs belong on it

Marketing KPIs (key performance indicators) are the few marketing performance indicators that tell you whether a plan is working. A marketing KPI dashboard shows them. The third column of the table below says whether you can add each one up across days and campaigns, or must recalculate it from totals.

KPI Formula Combining days or campaigns Usual source Trap
SpendSum of media costAddAd platformsSay whether fees and tool costs are included
Clicks, impressionsCountsAddAd platforms, Search ConsoleIn the sample, email impressions are emails delivered; keep them out of ad reach totals
CTRClicks ÷ impressionsRecompute from totalsAd platformsAveraging daily CTRs gives a quiet day the same weight as a busy one
CPCSpend ÷ clicksRecompute from totalsAd platformsLeave unpaid channels out
SessionsCountAddWeb analyticsSessions and platform clicks are different counts
UsersUnique people in the periodNever add daily users; pull the period totalWeb analyticsA visitor who comes on three days is one user
ConversionsCountAdd, one source per tileAd platform or web analyticsNever add the two sources together
Conversion rateConversions ÷ sessionsRecompute from totalsWeb analyticsPair sessions with analytics conversions
CPASpend ÷ conversionsRecompute from totalsBothAn incomplete last day has spend but few conversions
Conversion valueSum of valueAddWeb analytics or storeName the system the value comes from
ROASConversion value ÷ spendRecompute from totalsBothAveraging daily or row-level ROAS inflates it
CAC, MQLs, SQLs, pipelineCost ÷ new customers; lead-stage countsRecompute rates from totalsCRM or billing exportNot in ad data; define each stage

Download the sample and the KPI sheet

The sample has 900 rows, one per day and campaign: 12 campaigns across Paid search, Paid social, Email and Organic search, July 10 to October 7, 2026. It is synthetic sample data from a script with a fixed random seed, and it reproduces every number here. With the KPI sheet, it works as a digital marketing dashboard template.

marketing-dashboard-sample-90-days.csv  ·  marketing-dashboard-kpi-definitions.csv  ·  README

Six marketing dashboard examples

Each example gives the audience's question and its tiles. We built the first three in Anomaly AI from the sample; for the other three, we list the tiles only.

1. Executive dashboard (marketing director or CMO)

Question: Are we on target this month, and which channel is carrying it? Tiles: total spend, conversions, conversion value, paid CPA against target, paid ROAS against target, weekly spend and conversions, and a channel table.

Executive Marketing Monthly Review dashboard with the review period set to All complete days, showing total spend 104,510.21 USD, 3,450 analytics conversions, 383,342.96 USD conversion value, paid CPA 39.42 USD against a $40 target and paid ROAS 2.88 against a 3.0 target.
Example 1 shows the review-period filter and five KPI tiles, two of them against targets. First-party Anomaly screenshot; synthetic sample data.

Paid CPA ($39.42) meets its $40 target; paid ROAS (2.88) misses 3.0. The dashboard's channel table shows why: Paid search returns 3.65, Paid social 1.75.

Channel performance section, paid-channel targets CPA at most $40 and ROAS at least 3.0. Channel table: Paid search spend $62,148, 1,928 analytics conversions, $226,737 value, CTR 4.4%, CPA $32.23, ROAS 3.65; Paid social $42,362, 723, $74,127, CTR 1%, CPA $58.59, ROAS 1.75; Email and Organic search show $0.00 spend with CPA and ROAS left blank; total $104,510 spend, 3,450 conversions and $383,343 value.
This channel table explains example 1's ROAS miss. First-party Anomaly screenshot; synthetic sample data.

2. Paid media and campaign dashboard

Question: Which campaigns are wasting budget? This marketing campaign dashboard doubles as a marketing performance dashboard for the paid team. Tiles: spend, clicks, CTR, CPC, conversions, CPA and ROAS, a daily spend and CPA trend, a campaign table, and a channel filter.

Paid media performance dashboard with the paid channel filter set to Paid social, showing spend 42,361.97 USD, 36,255 clicks, CTR 1.0%, CPC 1.17 USD, 723 conversions recorded by web analytics, CPA 58.59 USD and ROAS 1.75.
Here example 2 is filtered to Paid social, and every tile recalculates. First-party Anomaly screenshot; synthetic sample data.

Filtered to Paid social, CPA is $58.59 and ROAS 1.75. Inside that channel, Retargeting returns 5.69 on $4,797 of spend while Prospecting returns 1.20 on $29,912.

3. Campaign launch dashboard

Question: Is the launch pulling its weight across channels? Tiles: spend, sessions, conversions, conversion value and paid ROAS since launch, daily sessions by channel, and a channel table.

Sessions since launch, September 8 to October 6, 2026, complete days only: a line chart of daily sessions by channel in which Email spikes to between about 600 and 850 sessions on three send days while Paid search and Paid social stay below about 400.
Example 3 shows daily sessions by channel since the launch: email sessions spike, and the paid channels carry the days in between. First-party Anomaly screenshot; synthetic sample data.

From September 8 to October 6 the launch spent $13,939.87 and recorded 298 conversions, with a paid ROAS of 1.95.

4. Email marketing dashboard

Question: Which sends drive visits and sales? Tiles: emails delivered, click rate (clicks ÷ delivered), sessions, conversions recorded by web analytics, conversions reported by the email tool, value per send, and a trend by send date. Email carries no media cost in the sample, so ROAS stays blank unless you add the tool's cost.

5. Organic search and content marketing dashboard

Question: Is organic search bringing the right visitors? An SEO and marketing dashboard puts search impressions, clicks and CTR next to sessions, conversions and value for blog pages versus product pages, plus a top-pages table. Search Console supplies the search side; you can connect Google Ads and Search Console to Anomaly directly.

6. Web analytics dashboard

Question: How are visitors arriving and what do they do? Tiles: sessions, users for the period, engaged sessions, conversions by source and medium, and top landing pages. If GA4 is your source, follow the guide to build a GA4 dashboard step by step.

Cross-channel marketing dashboard

A cross channel marketing dashboard puts channels side by side. It only works with one date window, one currency and one definition per KPI, and with each conversion figure labelled by its source. If you show GA4 conversions by channel, describe them as the conversion credit GA4 assigns under your attribution model, not as proof that a channel caused a sale.

Building the same view for several clients? For a client-by-client procedure, see repeatable client dashboards from messy CSVs.

How to build one, step by step

These six steps rebuild example 1 from the sample.

Step 1: Name the audience and the decision

Build one dashboard per audience. Write one sentence for each: who opens it, how often, and what they decide. "The marketing director checks monthly whether paid media hit its CPA and ROAS targets" leads to five tiles. "Show everything" leads to a wall of tiles.

Keep it to five to eight tiles, and give each KPI a target or a comparison period.

Step 2: Write each KPI's definition

For every tile, write the formula, the source column and the rule for combining days. The KPI sheet in the download lists fourteen KPIs, one per row. Decide now which conversion source each tile uses.

Calculation panel for the Paid ROAS tile, target 3.0 or more, on the Meaning tab: it divides total conversion value by total cost for Paid search and Paid social combined, using a ratio of totals rather than an average of row ROAS; value equals the sum of v over the sum of c, where v is conversion_value_usd and c is cost_usd for Paid search and Paid social before 2026-10-07.
The Paid ROAS tile's Meaning tab shows the formula, the terms and the ratio-of-totals rule. First-party Anomaly screenshot; synthetic sample data.

How it was checked: our own script recomputed paid ROAS from the CSV as 2.8788, the value this definition returns.

Step 3: Bring the data in

You can upload an Excel or CSV file up to 1 GB or use a Google Sheet. You can also connect sources directly: GA4, Google Ads, Meta and TikTok ads, Search Console, HubSpot, Shopify, BigQuery, MySQL and Snowflake. Ad networks without a connector can come in as a CSV or Sheets export. We uploaded the sample CSV.

Anomaly import screen with the Add Files drag-and-drop upload area above tiles for Google Sheet, Google Analytics 4 and Google Search Console.
On the import screen you upload a file or connect a source (the first row of sources is shown). First-party Anomaly screenshot; synthetic sample data.
Imported sample table showing 50 of 900 rows: the first nine rows with the columns date, channel, campaign, impressions, clicks, cost_usd and sessions; the table scrolls sideways to platform_conversions, analytics_conversions and conversion_value_usd.
This is the sample after upload: 900 rows, with the first seven of its ten columns in view. First-party Anomaly screenshot; synthetic sample data.

An uploaded file is a snapshot. To refresh it, upload the new export, or connect the workbook from OneDrive so it can be refreshed without re-uploading. Connected sources refresh before a scheduled report runs. If your monthly dashboard lives in Excel, see how to build and refresh a monthly dashboard from Excel.

Step 4: Ask for the dashboard

Anomaly AI, an AI data analyst for large datasets and spreadsheets, builds the dashboard from your file or connected sources, and every number comes from a stored query you can check. Put the audience, the tiles, the targets and the rules from step 2 into one request:

Prompt used for example 1

Build an executive marketing dashboard from the table marketing_dashboard_sample_90_days for a monthly review with the marketing director. Use complete days only: exclude 2026-10-07, which is a partial export. Show: total spend; conversions (use analytics_conversions and say so in the tile title); CPA and ROAS for the paid channels (Paid search and Paid social) against targets of CPA $40 and ROAS 3.0; total conversion value; a weekly trend of spend and conversions; and spend, conversions and ROAS by channel. Calculate CTR, CPA and ROAS as ratios of totals (for example SUM(conversion_value_usd) / SUM(cost_usd)), never as averages of daily or row-level values.

That request produced example 1: five KPI tiles, two weekly charts, a channel table and a notes block recording each rule. The table first showed unrounded ROAS and $0.00 CPA for unpaid channels; one follow-up request fixed both.

Step 5: Check the numbers before you trust them

Open the calculation behind the tiles that matter most. The query should match the definition you wrote in step 2: the right columns, the right filter and a ratio of totals. The final AND TRUE in the query is the review-period filter; here it is set to include all complete days.

Calculation panel for the Paid ROAS tile on the Code tab: SELECT SUM of conversion_value_usd divided by NULLIF of SUM of cost_usd, 0, FROM marketing_dashboard_sample_90_days WHERE the date is before 2026-10-07 AND the channel is Paid search or Paid social AND TRUE.
The Paid ROAS tile's Code tab divides summed value by summed cost. First-party Anomaly screenshot; synthetic sample data.

How it was checked: every headline number in examples 1 to 3 matched our independent script at the precision shown. Then run the four checks below.

Step 6: Filter, share, export or schedule

Add filters for the questions people ask most, such as channel or month. Then share the dashboard by a link, with or without an access code, or export it as PDF, PowerPoint, Word or Excel.

Access section of the share settings with two options, Anyone with the link (viewers with the link can view) and Password protected (requires an access code to view); Password protected is selected.
The share settings let you choose anyone with the link, or an access code. First-party Anomaly screenshot; synthetic sample data.
Top of page one of the PDF export of the executive dashboard in Anomaly's file preview: the title Executive Marketing Monthly Review, the five KPI values 104,510.21 USD, 3,450, 383,342.96 USD, 39.42 USD and 2.88, and the Weekly performance section heading.
This is the PDF export of example 1, top of page one, in Anomaly's file preview. First-party Anomaly screenshot; synthetic sample data.

A schedule sends the dashboard daily, weekly or monthly: as a PDF by email, to a Slack channel, or both. Set it to "Only when something happens" and, on every scheduled run, it checks for the change, threshold or exception you describe, and sends the report only when it is met. Scheduling is included from the Pro plan at $25 a month (pricing checked October 8, 2026). To set one up from scratch, follow the weekly marketing report tutorial.

New Schedule form: a weekly paid media check of paid CPA and paid ROAS for the last 7 complete days against the previous 7, set to run every Monday at 9am and to send Only when something happens, with the condition: paid CPA for the last 7 complete days is above $40, or paid ROAS is below 3.0.
The New Schedule form sends the report only if the condition is met, checked on each weekly run. First-party Anomaly screenshot; synthetic sample data.

When we saved it, Anomaly warned that an uploaded file keeps its data until you upload a new one. It also asked which weekday should close each week, so no partial day slips in. If you work in Google Sheets, see how to send a weekly PDF from Google Sheets, or turn an upload into a client-ready PDF.

Create an account and rebuild this dashboard from the sample

Four checks before you share it

Each check catches a number that looks fine but is wrong. We ran checks 1 and 3 in Anomaly on the sample; checks 2 and 4 follow Google's documentation.

1. Recompute ratios from totals

ROAS, CPA, CTR and conversion rate must be recalculated from summed totals. Averaging a ratio column gives a small row the same weight as a large one.

Aggregation comparison on the Checks before sharing dashboard, Paid search and Paid social combined on complete days: ROAS 2.8788 as a ratio of totals, 2.9978 as an average of daily ROAS and 4.0335 as an average of row ROAS; (2) is 4.13% and (3) is 40.11% above (1).
The same data gives three ROAS figures; this shows how far each strays from the ratio of totals. First-party Anomaly screenshot; synthetic sample data.

On the sample, paid ROAS from totals is 2.88. Averaging daily ROAS gives 3.00 (4.1% higher). Averaging a column with one ROAS per campaign per day gives 4.03 (40.1% higher). In that column average, a retargeting row with $14.03 of spend that returned 26 times its cost counts as much as a $500 row. The gap is specific to this illustrative sample and grows with how much spend varies.

2. Never add daily users

GA4 defines total users as the number of unique users who triggered an event "in the specified date range" (GA4 Help, verified October 8, 2026). A person who visits on Monday and Friday is one user that week but appears in two daily rows. Pull the period's user count from the source instead of summing days.

3. Drop the incomplete last day

A morning export includes part of today. In the sample, October 7 was exported mid-morning. It holds $641 of spend, about 45% of the daily average for the week before, but only 5 conversions, against 55 on October 6.

Two daily charts from September 24 to October 7: analytics conversions fall from 55 on October 6 to about 5 on the partial October 7, and paid CPA jumps from about 40 USD to about 160 USD on the same partial day.
Daily conversions and paid CPA: the partial day looks like a collapse. First-party Anomaly screenshot; synthetic sample data.
Week-over-week analytics conversions table, partial versus complete days: complete days only, 271 current against 307 previous conversions, -11.73%; including the partial day, 236 against 306, -22.88%.
Week over week, the partial day roughly doubles the drop. First-party Anomaly screenshot; synthetic sample data.

Paid CPA on the partial day is $160.28, against $41.69 on October 6. Week over week, conversions fall 11.73% on complete days but 22.88% when the partial day is included. Cut the date range at the last complete day, or label that day provisional.

4. Label which system each conversion figure comes from

Ad platforms and web analytics count conversions differently. Google Ads, for example, reports its primary conversion columns "based on the time of the click, not the time of the conversion" (Google Ads Help, verified October 8, 2026). In the sample, Paid social reports 1,054 conversions on the platform and 723 in web analytics.

Campaign results table on the Paid media performance dashboard, filtered to Paid social, with separate columns for ad platform conversions and web analytics conversions: Social - Prospecting spend $29,912, 533 and 360, CPA $83.09, ROAS 1.20; Social - Fall launch $7,653, 154 and 103, $74.30, 1.43; Social - Retargeting $4,797, 367 and 260, $18.45, 5.69; totals $42,362, 1054 and 723.
These campaign results have two conversion columns, each labelled with its source system. First-party Anomaly screenshot; synthetic sample data.
  1. Recompute ratios from totals; never average them across days or rows.
  2. Pull users for the whole period from the source.
  3. Make sure the last day is complete, or label it provisional.
  4. Label every conversion figure with its source system.

Where to build it

The table compares categories of marketing dashboard tools and software, not products, and is not ranked. We make Anomaly AI; it appears as one example in one row, described with the same columns.

Category Good when Watch out for Upkeep Cost type
Spreadsheet
Excel, Google Sheets		One or two sources and a small teamManual pasting and averaged ratio columnsHigh: re-export each periodIncluded in office software
Free report builder
e.g. Data Studio		Mostly Google sources such as GA4, Ads and SheetsOther sources need a community connector or a sheet in betweenMediumFree tool; some community connectors cost money
BI platform
e.g. Power BI, Tableau		An analyst or data team maintains a data modelSetup time and someone to own the modelMedium to highPer-user licences
Marketing reporting toolMany ad and social sources, or reports for clientsFixed templates and metric definitions chosen by the vendorLowSubscription
AI data analyst
e.g. a chat assistant, Anomaly AI		You can describe the dashboard in plain words, or the files are largeCheck each calculation; a one-off chat answer does not refreshLow to mediumSubscription or credits

How we wrote this table

The Data Studio details come from Google's data sources documentation, checked on October 8, 2026, including that community connectors may cost money. The other rows describe each category in general terms. Only Anomaly was run hands-on, on the sample. There are no scores and no ranking.

If the free report builder fits, here is how to build it in Looker Studio, which Google renamed Data Studio in April 2026. To choose a specific product, compare AI tools for marketing data analysis or see client reporting tools for agencies.

Why dashboards go unread

Marketing dashboards go unread for fixable reasons:

  • Vanity metrics. Impressions rise while the business question goes unanswered.
  • No owner. If nobody is named to act on a tile, nobody does.
  • Too many tiles. Every extra tile makes the number a reader came for harder to find.
  • No date on the data. Show the range and the last refresh so stale numbers are obvious.
  • Unlabelled sources. Two conversion counts turn the meeting into a debate.
  • No target. A number with nothing to compare against prompts no decision.

Frequently asked questions

What should a marketing dashboard include?

Spend, results and efficiency for the channels its audience owns, each against a target. In the examples above that means spend, conversions, CPA, conversion value and ROAS, a trend and a channel table: five to eight tiles.

How often should a marketing dashboard update?

Match it to the decision: daily for active paid campaigns, weekly for channel reviews, monthly for executives. Connected sources refresh before a scheduled report; an uploaded file stays as it is until you upload a new export.

Can I build a marketing dashboard in Excel or Google Sheets?

Yes, for one or two sources. Keep raw data on its own tab, compute ratios from summed totals, and date each refresh. Expect to re-export and paste the data every period.

Which KPIs should executives see compared with channel owners?

Executives need totals against targets: spend, conversions, CPA, ROAS and conversion value. Channel owners need the levers behind them: CTR, CPC, conversion rate and campaign-level CPA and ROAS.

Do I need a data warehouse for a marketing dashboard?

No. Every dashboard in this guide was built from one uploaded CSV. Exports, or direct connections to GA4, Search Console and ad accounts such as Google Ads, Meta and TikTok, are enough to start; a warehouse helps when many sources must be joined at scale.

How do I make a marketing dashboard for clients?

Use one layout and one set of KPI definitions for every client and swap only the data. Share each view by a code-protected link or a scheduled PDF, with the date range and conversion source labelled.

How we made this guide

We built the synthetic sample with a script that uses a fixed random seed, and ran it in Anomaly AI on October 8, 2026, in a project that held only this data. A separate script recomputed every number on this page from the CSV. Drafting and the product run were AI-assisted. Abhinav Pandey checks every number, caption and page detail against the sample before the page is released.

Ash Rai

Ash Rai

Technical Product Manager, Data & Engineering

Ash Rai is a Technical Product Manager with 5+ years of experience building AI and data engineering products, cloud and B2B SaaS products at early- and growth-stage startups. She studied Computer Science at IIT Delhi and Computer Science at the Max Planck Institute for Informatics, and has led data, platform and AI initiatives across fintech and developer tooling.

LinkedIn

Related Articles

Continue exploring AI data analysis with these related insights and guides.

View All Articles
Microsoft for Startups
UK GDPR compliantICO registered, reference ZC104879SOC 2 in progress