Revenue drop root cause analysis: a 7-step playbook for Stripe data
Why did revenue drop overnight? A revenue drop root cause analysis works best as an order of elimination. Rule out the cheap, boring causes first: a day that has not closed, payments that failed, tracking that broke. Then look at channels, campaigns and deploys, and write down what you found.
- Check first
- Is the day closed, in the right time zone, on the charge date? Did payment failures rise?
- Then
- Did tracking break? Which slice, channel or campaign holds the drop?
- Last
- What changed just before it started: a deploy, a plugin, an ad or a price?
This playbook is for whoever gets the message at 9am that yesterday looks wrong. Each step either rules a cause out or narrows where to look next. Stop as soon as one step explains the whole drop.
Step 1: Confirm the drop is real, not a reporting lag
Before you look for a cause, make sure there is something to explain. This is the fastest step to rule out, so do it first.
- Is the day finished? Compare the same hours. Today at 9am against all of last Tuesday will always look like a drop.
- Which time zone closes the day? If the report and your business use different time zones, part of one day lands in the next.
- Charge date or payout date? Stripe records each charge as it happens. Payouts land on a schedule days later. A low payout is not a low sales day.
- Gross or net? One large refund or dispute can pull net revenue down on a day when sales were normal.
- Is the baseline fair? Compare with the same weekday last week, not with yesterday. For subscriptions, check how many renewals were due. A day with fewer renewals due is lower by design.
When Stripe and your other reports disagree
A Stripe revenue discrepancy with GA4 or an ad platform is often timing, not money. Google says standard GA4 reports can take 24 to 48 hours to finish processing. Google Ads credits a conversion to the date of the click, so yesterday can keep rising for days. If the drop shows only in those tools, check Stripe. It is the closest record to the money.
Step 2: Segment by payment method, geography, and plan
Once the drop is real, find where it sits. A drop inside one slice points to a cause in that slice. A drop spread across every slice points to something everyone passes through.
| Where the drop sits | Look at first |
|---|---|
| One payment method or card brand | That method's setup, the provider behind it, and its decline codes. |
| One country or currency | Local holidays, currency or tax changes, and local payment methods. |
| Renewals only | Failed renewals, retries, and how many renewals were due that day. |
| New customers only | Traffic, the pricing page, the checkout and any recent price change. |
| One plan or product | Changes to that plan, its page, its stock or its price. |
| Every slice at once | Something everyone passes through: the site, the checkout or the payment flow. |
Compare each slice with its own baseline, not its share of today. A slice that is always small can look like the cause just because it fell to zero.
Step 3: Check payment failure and decline rates first
Payment failures are the cheapest check, and Stripe holds all of it. It records each failed payment as it happens, with a decline code. Compare today's failure rate with the same weekday baseline, then split the failures by type:
- The bank said no. A jump in one decline code from one country or card brand points to the card issuer, not to you.
- Your own fraud rules blocked it. A jump in blocked payments points to a rule that changed or one that is now too strict.
- The request failed. A jump in errors points to your checkout code or settings. Go straight to step 6 and look for a deploy.
For subscriptions, read failed renewals on their own. If you retry failed renewals, some of today's revenue may come back over the next few days. Count it as late, not lost, until the retries finish. Read refund reasons too, in case a spike in refunds is the real story.
Step 4: Rule out tracking and tag breakage
Stripe knows the payment, the amount and the customer's email. It has no record of the ad, search or email that brought the customer in. So use it as the referee.
- Stripe is steady, but your dashboard dropped. That is tracking, not revenue. Check whether the purchase event still fires, and whether a tag, a cookie banner or the thank-you page changed.
- Orders arrive with no source. If sales suddenly show as direct, the campaign tags were lost on the way. The money is fine. The credit is not.
- Stripe dropped, but visits held. People still came and did not pay. The leak is at the checkout or the payment step.
On a WordPress site, check each page's modified date and the plugin updates around the drop. A plugin update can break a checkout or a tag without anyone touching the page.
Step 5: Isolate channel and campaign contribution
Only now look at marketing. Earlier steps rule out the causes that would make channel numbers wrong. To tie Stripe charges to channels, match each charge to the order or the contact by email, then to the source of the visit. Shopify saves the landing page and referring site of the session that bought.
- Read the Google Ads change history for bids, budgets and ads around the drop.
- Check Meta for creative changes and delivery status by ad. A paused or rejected ad can stop a channel overnight.
- Check Search Console for the date a page's position moved.
- Check the email calendar. If last week had a send and this week did not, a normal day looks like a drop.
Meta reports conversions under its own attribution setting, seven days after a click and one day after a view by default. Do not treat yesterday's ad-platform numbers as final.
Step 6: Correlate with deploys and support ticket spikes
Put every change on one timeline, in one time zone. Root cause analysis needs timestamps precise enough to put events in order. The start time of the drop matters more than its size: look for what changed just before it.
- Deploys. Find the commit and deploy closest to when revenue moved. Stripe and GitHub share no customer ID, so line them up by time.
- Support tickets. Look at ticket tags and first-reply times around the drop. A run of tickets about paying or checking out confirms a break. No tickets does not rule one out, because many customers leave without writing in.
- Team chat. Incident threads show when the team noticed and what they decided.
A change next to the drop is a suspect, not a verdict. Test it. Check whether the drop sits in the slice the change touched, or roll the change back and watch.
Step 7: Write the incident note so it compounds
The investigation is only half the value. The note you write after it is what makes the next one faster. Keep every note in one place, in the same shape:
- What moved, by how much in dollars, and when it started and ended.
- The slices it hit, and the slices it did not.
- The cause, and the evidence that shows it.
- What you ruled out, and how you ruled it out.
- What fixed it, and how much revenue came back.
- What would have caught it sooner.
Tag each note with the kind of cause: reporting, payments, tracking, marketing or product. After a few incidents, you will see which steps find your problems, and you can move them up the list. Add each incident to your list of known events too, so your alerts and forecasts do not learn it as normal.
askotter reads your accounts side by side, so it can test each explanation against your connected accounts and find the cause. You get a ranked list of likely causes, each with the evidence behind it. Stripe, GA4, Google Ads, Meta Ads, Shopify and GitHub connect today. Zendesk and Slack connections are on the integration roadmap. See Stripe + GitHub root cause analysis, Stripe + Google Analytics root cause analysis and Stripe + Zendesk root cause analysis.