Marketing

Yesterday's numbers, already checked

One brief every morning with yesterday's spend, ROAS and the misses that need you — each number next to the table it came from.

Join the waitlist See what it produces

Before standup

Check your ads daily

Who it is for: Performance and growth leads whose ad numbers already land in a warehouse table.
What starts it: Every morning at 07:00 in your workspace time zone, or on demand from the ticket you are already looking at; On a schedule

Coming soonCheck your ads daily is built, but not ready yet, so it cannot be turned on today. In Aestus you can ask to be told when it is ready.

A morning ads check is mostly assembly. Two ad managers, an export tab, a sheet whose formulas nobody has re-derived in a year, and twenty minutes later you have a number you would not want to defend. The work that matters — deciding what to do about the campaign that slipped — starts after all of that, on the least alert part of the morning.

So the first thing to say about this one is what it does not do, because it is the thing a page like this usually hides. It does not read an ad account. Aestus has no Google Ads connection, no Meta connection, no TikTok connection and no LinkedIn connection. There is no vetted tool list for any of them, and a use case that declared one would put a connect button in front of an integration that does not exist — which fails at the exact moment somebody trusts it. It reads your warehouse instead: the campaign rows your team already exports there. That is not a smaller version of the idea. It is the only shape in which one brief can cover Meta and TikTok at all, because a per-platform connection covers one platform and your export covers whichever ones you point it at.

The second thing is where the numbers come from. The step that reads your warehouse is allowed to report exactly one kind of thing: whole numbers and money columns that a query returned. It reports no ROAS, no CPA, no CTR, no delta and no total — and the reply it has to produce has no field to put one in. Every metric you read was computed afterwards, by software, from figures you can go and check. That is a smaller claim than "the model is careful with arithmetic", and it is the only one worth making, because a wrong ROAS reads exactly like a right one and the person who finds out has already moved a budget.

A campaign with no conversions shows n/a for its cost per conversion. Not zero, and not a missing column. This is a small detail that decides whether the section is worth reading: a cost of zero is the most attractive number on the page, and it is the one number there that means nothing at all.

The third thing is disclosure. Every brief carries a Not connected section whether or not there is anything in it, and a platform your parameters name that no table covers is listed there with the reason. Its spend is called unknown, never zero. A section that only appears when there is something to say is a section that disappears on the morning it mattered, and an absence is the one thing a reader cannot see for themselves.

What reaches your board is one ticket per miss that needs a person, keyed by the platform, the campaign and the rule it broke. Tomorrow's run rebuilds that key and finds the ticket rather than opening a second one, so a campaign that misses its target all week is one ticket with five updates. The brief names what missed and by how much, against which target, and which setting that target came from. It does not tell you to pause anything. Nothing in the run can reach an account to check whether pausing worked, and advice nobody can stand behind is worse than no advice in a document that otherwise checks its own sums.

What you get

The artifact this use case produces, in the product view you would read it in.

Example — fictional team and data. Not a live run or customer result.

Worked example

document

Daily ads brief — 2026-09-10

Produced by: Run EXAMPLE-ADS-01 · warehouse read-only · document artifact · demo values

Yesterday: Spend 12,845.30 USD. Revenue 42,370.95 USD. ROAS 3.30 against a target of 3.00. 390 conversions at a blended CPA of 32.94 USD against a ceiling of 45.00. CTR 2.96% over 1,845,200 impressions. 2 of 5 campaigns need a person. 1 platform is not connected.
Act now: 1. **Google Ads — Non-brand Search: ROAS 1.74 against 3.00** (`targetRoas`), 42% below. Source: warehouse `ads.google_campaign_daily`, date column `report_date`. 2. **Google Ads — Non-brand Search: CPA 71.71 against 45.00** (`targetCpa`), 59% above. Source: warehouse `ads.google_campaign_daily`, date column `report_date`. 3. **Meta — Advantage+ Shopping: ROAS 2.10 against 3.00** (`targetRoas`), 30% below. Source: warehouse `ads.meta_campaign_daily`, date column `report_date`.
Watch: - Total spend 12,845.30 USD is 92% of the 14,000.00 USD daily cap (`dailySpendCap`).
Clear: - Google Ads — Brand Search: ROAS 7.09, CPA 17.36 USD. - Google Ads — Performance Max Retail: ROAS 3.92, CPA 30.59 USD. - Meta — Retargeting DPA: ROAS 6.00, CPA 14.59 USD.
By campaign: | Platform | Campaign | Spend | Revenue | ROAS | Conv. | CPA | CTR | Source | | --- | --- | --- | --- | --- | --- | --- | --- | --- | | Google Ads | Brand Search | 1,284.40 | 9,112.60 | 7.09 | 74 | 17.36 | 7.55% | ads.google_campaign_daily, report_date | | Google Ads | Non-brand Search | 2,940.15 | 5,116.80 | 1.74 | 41 | 71.71 | 2.72% | ads.google_campaign_daily, report_date | | Google Ads | Performance Max Retail | 3,610.00 | 14,152.25 | 3.92 | 118 | 30.59 | 2.42% | ads.google_campaign_daily, report_date | | Meta | Advantage+ Shopping | 4,120.75 | 8,650.40 | 2.10 | 96 | 42.92 | 1.46% | ads.meta_campaign_daily, report_date | | Meta | Retargeting DPA | 890.00 | 5,338.90 | 6.00 | 61 | 14.59 | 2.16% | ads.meta_campaign_daily, report_date | | **Total** | | **12,845.30** | **42,370.95** | **3.30** | **390** | **32.94** | **2.96%** | | ````aestus-chart id: chart_ad5da11d0a11 kind: bar title: "ROAS against target by campaign — 2026-09-10" categories: ["Non-brand Search","Advantage+ Shopping","Performance Max Retail","Retargeting DPA","Brand Search"] --- - { id: series_0a5d1a11, name: "ROAS", values: [1.74,2.1,3.92,6,7.09] } - { id: series_7a4e0e71, name: "Target", values: [3,3,3,3,3] } ````
Not connected: This brief does not read any advertising platform. Aestus has no Google Ads, Meta, TikTok or LinkedIn connection, so nothing here is fetched from an ad account. Every number comes from your own warehouse tables — whatever your team already exports there. A platform with no table configured is named under Not connected: its spend is unknown, never zero. - TikTok Ads — no warehouse table is configured for this platform. Its spend is unknown, not zero.
Caveats: - Conversions and conversion value keep being attributed after a day closes, so yesterday's figures can still move. This brief reports what the table held when it was read. - No currency is converted. Rows are reported in the currency the table stores them in, and a mixed-currency total is refused rather than summed. - The brief reads your warehouse, not an ad account. A campaign your export does not cover is invisible to it, and a stale export looks exactly like a quiet day.
Sources: - Warehouse table `ads.google_campaign_daily (date column `report_date`)`. - Warehouse table `ads.meta_campaign_daily (date column `report_date`)`. - Query on `ads.google_campaign_daily` returning 3 rows: `SELECT campaign, spend, revenue, conversions, clicks, impressions FROM ads.google_campaign_daily WHERE report_date = $1` - Query on `ads.meta_campaign_daily` returning 2 rows: `SELECT campaign, spend, revenue, conversions, clicks, impressions FROM ads.meta_campaign_daily WHERE report_date = $1` - Prior briefs read for streaks: Daily ads brief — 2026-09-08, Daily ads brief — 2026-09-09.

How it runs

Each step is one node in the workflow graph the interpreter runs. Nothing here is a prompt chain hidden in a route handler.

Workflow steps

01Read yesterday
Read-only SQL against the warehouse tables you name pulls one row per campaign for the day that just closed: spend, revenue, conversions, clicks and impressions, exactly as your export stores them. Nothing is written back, and no ad account is touched.
02Count in code, not in the model
The reading step returns raw numbers and nothing else. ROAS, CPA, CTR and every total are computed from them by software you can check. The reply it must produce has no field for a rate, so there is nowhere to put one that no query produced.
03Rank against your targets
A target missed by more than ten percent is Act now; a miss inside ten percent is Watch; everything else is Clear. Each flag names the metric, its value, the target and the setting the target came from, so you can go and change it.
04Check its own arithmetic
Before anything is published, a second check recomputes every ROAS, CPA and CTR from that row's own raw numbers, re-sums the totals, and confirms every platform it could not read is named. A brief that fails twice is not published, and the run records why.
05Publish, post and file
The brief becomes a document on the day's ticket, goes to your channel once you approve the send, and each miss that needs a person becomes one ticket — updated on each day it persists rather than filed again.

By hand, and on Aestus

The same job, on either side.

By hand
Open two ad managers and a spreadsheet, one after another, before anyone has spoken.
On Aestus
One brief, every table read in the same run, waiting when you sit down.
By hand
Retype yesterday's numbers into a sheet whose formulas nobody has checked since the person who wrote them left.
On Aestus
ROAS, CPA and CTR computed from the raw rows, then recomputed by a second check before publishing.
By hand
A number in the standup with no record of where it came from.
On Aestus
Every table row ends with the warehouse table and the date column behind it.
By hand
A platform nobody exported looks exactly like a platform that spent nothing.
On Aestus
A source it could not read is named under Not connected, with the reason, and its spend is called unknown rather than zero.
By hand
The same overspend mentioned in the channel five mornings running.
On Aestus
One ticket per miss, updated each day it persists, in the project you chose.
By hand
“Is this the third bad day or the first?” answered from memory.
On Aestus
A streak is written down only when the brief can name the earlier briefs that show it.
By hand
The check happens on the mornings somebody remembers.
On Aestus
A run row every morning, with its cost, duration and outcome.

What Aestus records

Every run is auditable after the fact, which is what makes handing the job over survivable.

The brief
A document artifact on the day's ticket, with the eight sections in the same order every morning, the per-campaign table and the ROAS-against-target chart.
One ticket per miss
In the project you choose, keyed by platform, campaign and rule, so a miss that lasts a week is one ticket with a week of updates rather than seven tickets.
The run
Cost, tokens, latency and failure classification, the same as every other dispatch in Aestus. A brief that failed its own arithmetic check twice records that instead of publishing.

What it connects to

Only the tools this workflow actually uses. Each one has a setup guide.

slackOptional

Questions teams ask

Does it connect to Google Ads?

No. Aestus has no Google Ads, Meta, TikTok or LinkedIn connection, so nothing here is fetched from an ad account. It reads the campaign rows your team already exports into your own warehouse, which is also the only way one brief can cover every platform at once. A platform with no table is listed as not connected rather than left out.

Can it change my campaigns?

No, and not because it was told not to. There is no tool in the workflow that can reach an ad account, so no bid, budget, status or creative is reachable from it. It names what missed and by how much; the spending decision stays with you.

Do the numbers move after the brief is written?

They can. Both platforms keep attributing conversions after a day closes, so the brief reports what your table held when it was read and says so every time.

What if it gets the arithmetic wrong?

A second check recomputes every ROAS, CPA and CTR from that row's own raw numbers before publishing. A brief that fails twice is not published at all, and the run records the reason.

What happens to a campaign with no conversions?

Its cost per conversion reads n/a, never zero and never a dropped column. A zero there would say the opposite of the truth in the one place somebody would act on it.

All use cases How access is governed

Sources

lib/domain/use-cases/check-your-ads-daily/reader.ts
lib/domain/warehouse/sql-policy.ts
content/integrations/slack.md
content/integrations/google-ads.md

Hand over Check your ads daily.

Tell us how your team does this today and we will set it up with you.

Talk to us

Check your ads daily

Who it is for

Performance and growth leads whose ad numbers already land in a warehouse table.

What starts it

Every morning at 07:00 in your workspace time zone, or on demand from the ticket you are already looking at

On a schedule

Coming soonCheck your ads daily is built, but not ready yet, so it cannot be turned on today. In Aestus you can ask to be told when it is ready.

Daily ads brief — 2026-09-10

Produced by: Run EXAMPLE-ADS-01 · warehouse read-only · document artifact · demo values

Yesterday: Spend 12,845.30 USD. Revenue 42,370.95 USD. ROAS 3.30 against a target of 3.00. 390 conversions at a blended CPA of 32.94 USD against a ceiling of 45.00. CTR 2.96% over 1,845,200 impressions. 2 of 5 campaigns need a person. 1 platform is not connected.
Act now: 1. **Google Ads — Non-brand Search: ROAS 1.74 against 3.00** (`targetRoas`), 42% below. Source: warehouse `ads.google_campaign_daily`, date column `report_date`. 2. **Google Ads — Non-brand Search: CPA 71.71 against 45.00** (`targetCpa`), 59% above. Source: warehouse `ads.google_campaign_daily`, date column `report_date`. 3. **Meta — Advantage+ Shopping: ROAS 2.10 against 3.00** (`targetRoas`), 30% below. Source: warehouse `ads.meta_campaign_daily`, date column `report_date`.
Watch: - Total spend 12,845.30 USD is 92% of the 14,000.00 USD daily cap (`dailySpendCap`).
Clear: - Google Ads — Brand Search: ROAS 7.09, CPA 17.36 USD. - Google Ads — Performance Max Retail: ROAS 3.92, CPA 30.59 USD. - Meta — Retargeting DPA: ROAS 6.00, CPA 14.59 USD.
By campaign: | Platform | Campaign | Spend | Revenue | ROAS | Conv. | CPA | CTR | Source | | --- | --- | --- | --- | --- | --- | --- | --- | --- | | Google Ads | Brand Search | 1,284.40 | 9,112.60 | 7.09 | 74 | 17.36 | 7.55% | ads.google_campaign_daily, report_date | | Google Ads | Non-brand Search | 2,940.15 | 5,116.80 | 1.74 | 41 | 71.71 | 2.72% | ads.google_campaign_daily, report_date | | Google Ads | Performance Max Retail | 3,610.00 | 14,152.25 | 3.92 | 118 | 30.59 | 2.42% | ads.google_campaign_daily, report_date | | Meta | Advantage+ Shopping | 4,120.75 | 8,650.40 | 2.10 | 96 | 42.92 | 1.46% | ads.meta_campaign_daily, report_date | | Meta | Retargeting DPA | 890.00 | 5,338.90 | 6.00 | 61 | 14.59 | 2.16% | ads.meta_campaign_daily, report_date | | **Total** | | **12,845.30** | **42,370.95** | **3.30** | **390** | **32.94** | **2.96%** | | ````aestus-chart id: chart_ad5da11d0a11 kind: bar title: "ROAS against target by campaign — 2026-09-10" categories: ["Non-brand Search","Advantage+ Shopping","Performance Max Retail","Retargeting DPA","Brand Search"] --- - { id: series_0a5d1a11, name: "ROAS", values: [1.74,2.1,3.92,6,7.09] } - { id: series_7a4e0e71, name: "Target", values: [3,3,3,3,3] } ````
Not connected: This brief does not read any advertising platform. Aestus has no Google Ads, Meta, TikTok or LinkedIn connection, so nothing here is fetched from an ad account. Every number comes from your own warehouse tables — whatever your team already exports there. A platform with no table configured is named under Not connected: its spend is unknown, never zero. - TikTok Ads — no warehouse table is configured for this platform. Its spend is unknown, not zero.
Caveats: - Conversions and conversion value keep being attributed after a day closes, so yesterday's figures can still move. This brief reports what the table held when it was read. - No currency is converted. Rows are reported in the currency the table stores them in, and a mixed-currency total is refused rather than summed. - The brief reads your warehouse, not an ad account. A campaign your export does not cover is invisible to it, and a stale export looks exactly like a quiet day.
Sources: - Warehouse table `ads.google_campaign_daily (date column `report_date`)`. - Warehouse table `ads.meta_campaign_daily (date column `report_date`)`. - Query on `ads.google_campaign_daily` returning 3 rows: `SELECT campaign, spend, revenue, conversions, clicks, impressions FROM ads.google_campaign_daily WHERE report_date = $1` - Query on `ads.meta_campaign_daily` returning 2 rows: `SELECT campaign, spend, revenue, conversions, clicks, impressions FROM ads.meta_campaign_daily WHERE report_date = $1` - Prior briefs read for streaks: Daily ads brief — 2026-09-08, Daily ads brief — 2026-09-09.

Questions teams ask

Does it connect to Google Ads?

Can it change my campaigns?

Do the numbers move after the brief is written?

They can. Both platforms keep attributing conversions after a day closes, so the brief reports what your table held when it was read and says so every time.

What if it gets the arithmetic wrong?

A second check recomputes every ROAS, CPA and CTR from that row's own raw numbers before publishing. A brief that fails twice is not published at all, and the run records the reason.

What happens to a campaign with no conversions?

Its cost per conversion reads n/a, never zero and never a dropped column. A zero there would say the opposite of the truth in the one place somebody would act on it.

Sources

lib/domain/use-cases/check-your-ads-daily/reader.ts

lib/domain/warehouse/sql-policy.ts

content/integrations/slack.md

content/integrations/google-ads.md

Yesterday's numbers, already checked

Check your ads daily

What you get

Daily ads brief — 2026-09-10

How it runs

By hand, and on Aestus

What Aestus records

The brief

One ticket per miss

The run

What it connects to

Questions teams ask

Other use cases

Hand over Check your ads daily.

Yesterday's numbers, already checked

Check your ads daily

What you get

Daily ads brief — 2026-09-10

How it runs

By hand, and on Aestus

What Aestus records

The brief

One ticket per miss

The run

What it connects to

Questions teams ask

Other use cases

Hand over Check your ads daily.