A Meta Ads Dashboard: Spend, CPA and Creative

Meta rewards constant tinkering and punishes judging too early. A good dashboard shows spend, cost per result and ad-level performance with a rule for when a number counts.

A Meta Ads dashboard should show four things: spend against budget, cost per purchase and ROAS by campaign, cost per purchase by ad set with a minimum spend rule, and a table of ads by spend with their results. Add a click-through trend for your main ads to catch fatigue. The rest is optional.

Principles that apply to Meta specifically

Meta's learning system makes budget and audience changes slow to show their effect, and small samples are noisy. Three principles follow.

• Judge on enough spend. An ad set with one purchase on a small budget will rank first on cost per purchase. Set a spend floor, such as two to three times your target cost per purchase, and hide anything below it.
• Pair every ratio with its base. Cost per purchase beside purchases and spend.
• State the attribution setting. Purchases depend on it. A dashboard that does not say which setting it uses cannot be compared with last year's.

The views

Spend and pacing

Daily spend by campaign for the current month, with the month total against budget. Meta can overspend on a day within its daily budget rules, so you look at the month, not the single day.

Campaign efficiency

For each campaign: spend, purchases, purchase value, ROAS and cost per purchase, ordered by spend. ROAS is purchase value over spend summed per campaign. If you also run lead objectives, keep a separate table with leads and cost per lead, since a lead campaign has no purchase value to rank on.

Ad set efficiency

Cost per purchase by ad set for the last 30 days, hiding ad sets under your spend floor. This is the view that finds a weak audience inside a healthy campaign.

Ad-level results

Ads sorted by spend for the last full month: purchases, purchase value, cost per purchase and link click-through rate. Name your ads with the concept and format, for instance "UGC video, free shipping", so a pivot by name can group concepts later. Meta's own reporting does not tell you which concept an ad belongs to unless you put it in the name.

Fatigue trend

Link click-through rate and cost per link click over time for the biggest campaigns and ad sets (weekly) and the biggest ads (monthly, since the ad data is monthly). A steady slide with stable targeting usually means the audience has seen the ad enough. Compute the rate as total link clicks over total impressions per week, not an average of daily rates. The reasoning is in what your CTR, CPC and CPM columns are not telling you.

By hand versus by asking

By hand, this is a breakdown in Ads Manager exported to a spreadsheet each week, with pivot tables for each view. It works, and the setup in a spreadsheet dashboard applies. The recurring cost is the weekly export and the risk of a changed column or filter.

Getting the views from Quiriz

Quiriz's Meta Ads connector is in beta. It reads one ad account, read-only, for 13 months using the attribution setting in Ads Manager, refreshes daily and is included in paid plans. Datasets:

• Campaign performance (daily) for pacing and campaign efficiency.
• Ad set performance (daily) for audience efficiency.
• Ad performance (monthly, top 2,000 ads per month by spend) for the ad table.

Meta Ads spend, impressions, clicks (link clicks), conversions (purchases), conversion value, leads, ROAS, CPC, CTR and CPA are each defined once, with ratios computed from totals. Questions to ask:

• "How much did we spend on Meta Ads each month this year?"
• "Which ad sets had the lowest cost per purchase in the last 30 days, with at least a set amount of spend?"
• "Which Meta ads spent the most last month, and what did they return?"
• "How has link click-through rate changed by month for our top ads?"

Keep the questions you repeat as saved questions and rerun them after each refresh. The answer includes its table, so spend can be checked against Ads Manager in a minute.

An illustrative weekly review

A practical weekly review takes ten minutes. Check month-to-date spend against budget. Scan campaign cost per purchase for anything more than about a third above target on meaningful spend (a rule of thumb for this example, set your own). Open the ad set table, hide anything under the spend floor, and note the best and worst two. Open the ads table and mark any ad that has spent more than the floor with no purchases. Write one line per decision: pause, scale, wait or test. The line matters because next week you can check whether the decision worked, which is how a dashboard becomes a learning loop instead of a scoreboard.

What the dashboard cannot show

• The creative itself. You get ad names and numbers, not images, videos or copy. Creative analysis depends on naming conventions you control.
• Reach and frequency. They are not additive across days, so they are not included. The click-through trend is a proxy for fatigue, not a replacement.
• Placement, age, gender and country breakdowns are not read.
• Monthly ad grain. The ad dataset is by month, so a weekly ad-level trend is not available. Weekly trends work at campaign and ad set level.
• Top 2,000 ads per month by spend, so the ad total can sit below the campaign total.
• Purchases are as Meta counts them and will not match your store; see why.
• Facebook sign-ins expire after roughly 60 days, so you will be asked to reconnect.

Related reading: asking your Meta ad account questions, and why Facebook and Google ROAS do not compare. For agencies reporting this to clients, see Quiriz for agencies.

Frequently asked questions