Support for marketing platforms: "the numbers don't match" is handled by playbook

A martech platform user is a professional who writes to support when a report does not add up before a client meeting. They need not a polite reply but an analysis: where the discrepancy comes from, what to check, when it will be fixed. We build the first line for B2B platforms so that typical discrepancies are analysed by checklist and only real defects go to development — with reproduction.

What comes into a platform's support

• Connection and access: an ad account does not connect, a token expired, no rights to the needed account.
• Data discrepancies: the platform's figures do not match the ad account — the most frequent and most complex category.
• Reports and dashboards: wrong filters, wrong period, wrong currency, an empty report.
• Auto-rules and alerts: a rule did not fire or fired at the wrong time.
• Billing: limits, charges, invoices, plan changes.
• Roles and teams: who sees what, invitations, account transfers.

Discrepancy analysis by checklist

Most "the numbers don't match" requests are not a platform defect but a difference in source rules: attribution window, time zone, data update delay in the ad account, rounding, a campaign status filter. The first line goes through the checklist of these causes with the user before escalating. If the discrepancy is not explained by the checklist and reproduces across several periods — it is a defect, and it goes to development with data: account, period, metric, figures from both sides, screenshots.

1. Clarify the account, period, metric and attribution on both sides.
2. Check against the list of known causes: time zone, attribution window, data delay, filters, rounding.
3. If explained — reply with the explanation and a link to the knowledge base article.
4. If not — reproduce on a second period and escalate to development with data and severity.
5. Tell the user the deadline and notify after the fix.

Support that unloads development

In B2B platforms developers often answer users themselves because nobody else understands the integrations. That costs the team days of work a week. A first line that can read integration logs, check tokens and reproduce a discrepancy takes everything that needs no code change off development and passes the rest as a ticket that can be picked up immediately. Requests also become a source of regression cases for integration testing.

Channels, hours and tone

For B2B — email, in-product chat, Telegram and a Slack connect with key clients. Coverage hours — the business hours of the clients' markets, extended for the month-end reporting period when requests peak. Tone — professional and specific: no apologies instead of an answer, with figures and links to documentation. Russian, English and Uzbek in-house.