For agencies that grow D2C brands
How do you find out why a brand's numbers moved before the brand asks?
To diagnose performance, an agency checks what changed across a brand's ads, search, site and rivals, then ranks the likely causes by evidence. ai-dm does that for every brand, grades each cause by confidence and shows the platform records behind it, so your team explains the drop instead of guessing.
No plan to pick. Billed in USD and INR. Nothing changes without your approval.
records · meta ads ad-level report · ga4 purchase events · meta ad library
records · google ads search terms report · site crawl · change history
records · search console queries and pages · page indexing report
What do brands ask when a number drops?
The question always arrives before the answer. Here are four that agencies hear every week, answered without and with ai-dm.
Your brand asks
Why did our ROAS fall last week?
Someone exports three reports, lines them up in a sheet and offers a theory by Wednesday.
The likely causes were ready on Monday. Ranked by confidence, with the ad-level records behind each one and the causes that were ruled out.
Your brand asks
Why does GA4 show fewer sales than Meta?
The team picks the number that looks best, and the brand stops trusting both.
Both numbers, side by side, with how each is counted. Every figure links to the platform record it came from, so the gap has an explanation, not an excuse.
Your brand asks
Is our Google Ads budget being wasted?
The search terms report gets opened once a month, if the account manager has time.
Terms with spend and no conversions are flagged daily. Negative keywords are drafted for your team to approve, with the spend each one would have saved.
Your brand asks
Why are we getting fewer visitors from Google?
"Probably an algorithm update." Nobody checks which pages lost clicks, or why.
The pages and queries that dropped, named. Checked against the site crawl, last year's same week and the rivals that moved up.
No plan to pick. Add brands one at a time; each one's first report is built from its domain.
Why don't GA4, Meta and Google Ads agree on sales?
Because each one counts in its own way. Meta's default setting credits a sale up to 7 days after a click or 1 day after a view, Google Ads records a conversion against the day of the click, and GA4 shares credit across channels with its own model. All three can be right at once.
Counts purchases after a click or a view of the brand's ads.
record · ads manager · account 4412… · 7-day click, 1-day view
Counts conversions after an ad click, dated to the click.
record · google ads · customer 718-…-204 · conversions by click date
Counts purchases on the site, credited across all channels.
record · ga4 property 3209… · purchase key events
Hover or focus a card to see the record behind its number.
Illustrative example(login required) · Connect Meta, Google Ads and GA4 for one brand to start.
How does ai-dm work out why a number moved?
The same way a careful analyst would, every day, for every brand: start from the change, list everything else that moved, and keep only the causes the records support.
- 1
Spot the change
Each metric is measured against the brand's own baseline, so a real shift stands out from a quiet week.
roas vs 14-day baseline · −38%
- 2
List what else moved
Budgets, bids, creatives, search terms, the site, tracking, rivals and the season, for the same dates.
12 candidate causes · same 7 days
- 3
Grade the evidence
Each candidate is graded high, medium or low, and the ones the records contradict are ruled out.
2 likely · 1 ruled out · 9 not supported
- 4
Show the records
Every cause links to the platform record behind it, so your team can check it before telling the brand.
ad-level report · search terms · crawl
Readouts under each step: illustrative example.
A diagnosis usually starts with a daily change alert, and when the cause is on the brand's site, an SEO audit of the brand's site shows which pages broke.
Nothing changes in a brand's accounts until your team approves it.
What does diagnosis look like on a real brand?
Moolwan is our own D2C home décor brand from Bengaluru, run on ai-dm every day. Its weakest numbers are the ones ai-dm keeps raising.
A small budget against marketplaces. Moolwan spent ₹28,064 on Meta ads in 30 days, competing for the same shoppers as Amazon, Flipkart and Myntra. Its ROAS of 2.13× sits below a 4× e-commerce benchmark, and ai-dm flags it every day until it moves.
Strong and weak, side by side. The same brand drew 344,000 impressions in Google's AI results while its website health sits at 70%. Diagnosis means showing both, with the source of each number, so the next move is aimed at the real gap.
| Number | What it is | Source and date |
|---|---|---|
| 2.13× | Meta ROAS on ₹28,064 spend, below a 4× e-commerce benchmark | ai-dm · Paid ads · 30 days to 26 Sep 2026 |
| 171,489 | Meta ad impressions in the same 30 days | ai-dm · Paid ads · 30 days to 26 Sep 2026 |
| 3,895 | Clicks from those impressions | ai-dm · Paid ads · 30 days to 26 Sep 2026 |
| 344k | Impressions in Google's AI results | Search Console · Generative AI features · 25 Jun–23 Sep 2026 |
| 70% | Website health, with 1 critical and 10 high findings | ai-dm · 150-page crawl · Sep 2026 |
real data Moolwan's own accounts · bot traffic excluded
What do you want to diagnose first?
Each part of diagnosis has its own guide, with the checks to run and how ai-dm runs them for every brand you manage.
- any metric
Find why numbers moved
The likely causes of a drop or a jump, graded by confidence, with the records behind them.
See how to find the cause → - meta ads
Audit Meta ads
Find the ad sets, creatives and audiences wasting a brand's spend.
See the Meta ads audit → - google ads
Audit Google Ads
Search terms with spend and no conversions, match types and budgets, checked daily.
See the Google Ads audit → - linkedin · pinterest · x
Audit LinkedIn, Pinterest and X ads
The same checks for the smaller channels, so they get looked at too.
See the smaller-channel audit → - organic social
Analyse organic channel performance
Which posts and channels earn reach and clicks, and which only fill the calendar.
See organic channel analysis → - ga4 · meta · google ads
Reconcile GA4, Meta and Google Ads
Each platform's number side by side, with why they differ and the record behind each.
See the reconciliation → - search
Find keyword and content gaps
The searches rivals rank for and the brand doesn't, ranked by what they're worth.
See keyword and content gaps → - evidence
Trace every number to its source
Click any metric through to the raw platform record it came from.
See the evidence trail →
What would ai-dm find for your brand today?
Enter one brand's domain. ai-dm builds its first report from the domain alone: search, AI search visibility, competitors and website health. Connect its ad accounts and GA4 to diagnose paid performance.
Rather talk first? WhatsApp the founder
Questions agencies ask about diagnosing performance
What does it mean to diagnose marketing performance?
Diagnosing performance means finding out why a brand's numbers changed: which campaigns, pages, search terms or outside events moved, and how strong the evidence is for each. ai-dm does this every day for every brand you manage, against each brand's own baseline, and shows the platform records behind every likely cause.
Does ai-dm tell us which channel caused a sale?
No. ai-dm doesn't build attribution models, so it never claims that one channel caused a sale. It shows what changed, the likely causes graded by confidence and each platform's own numbers with their sources. Your team makes the call with the evidence in front of it.
Why do GA4 and Meta report different numbers for the same brand?
Because they count differently. Meta's default setting credits a purchase up to 7 days after a click or 1 day after a view, Google Ads records conversions against the day of the click, and GA4 shares credit across channels with its own model. ai-dm shows each figure side by side with how it was counted, instead of merging them.
Which ad platforms can ai-dm diagnose?
Meta Ads and Google Ads, plus LinkedIn, Pinterest and X once those accounts are connected. Search Console, GA4 and the brand's website add the organic and site side, so a drop in paid results can be checked against changes on the site and in search.
How confident is a diagnosis?
Each likely cause carries a grade of high, medium or low, based on how strongly the records support it, and causes the records contradict are marked as ruled out. Every cause links to the record behind it, so your team can check it before telling the brand.
How is ai-dm priced for an agency?
Billed in USD and INR. Per-brand plans or pay as you go. There is no plan to pick before you start: sign up, pay as you go, and add brands as you need them. Talk to the founder on WhatsApp if you want to plan a larger roster first.
Nothing goes live without your approval. A brand's data is deletable on request.
Win the why.™
Every number that moves has a reason. Dashboards show what happened; they rarely show why. Without the why, teams guess, act late and lose trust. ai-dm exists to win the why for every brand, every day: to find the reason behind every change, prove it with evidence, recommend the next move and check that it worked. Then it starts again, because the why keeps changing. When you win the why, you stop guessing and start growing. That is why we exist.
Gather. Analyse. Detect. Alert. Recommend. You act. Monitor. Learn and repeat. Always on.
Have the answer before your brand asks.