Audit and restructuring of an e-mail program for B2B SaaS: deliverability, segmentation, triggers and A-B tests in Customer.io

Initial situation: B2B SaaS for HR (subscription, deal cycle 2-6 weeks). Now mailings come from Customer.io, the sender domain is shared with transactional letters, some events come from Segment. The open rate in the database is 18-22%, clicks are 1-1.6%, noticeable failures in Gmail/Outlook, periodically letters end up in Promotions/Spam. Unsubscriptions increase after mass mailings. The email funnel is collected “as needed”, there are duplicates, there is no uniform logic of segments and frequency restrictions.

Technical environment and limitations: Customer.io + Segment (event-stream), product analytics in GA4, HubSpot CRM (read-only), sending from a dedicated subdomain (you can change DNS from Cloudflare). You cannot connect new paid services or buy new domains. You cannot change the texts of product letters (password reset, invoice) - only infrastructure and marketing chains. You cannot do an HTML redesign of emails - we work with current templates; spot edits are allowed that affect spam factors and click-through rates. Access to data is through guest roles, uploading personal data is prohibited, we work with anonymized reports/segments within systems.

What needs to be done (several steps):
1) Diagnostics of deliverability and reputation. Check SPF/DKIM/DMARC, alignment, return mail, tracking domain settings, presence/correctness of List-Unsubscribe. Collect a picture by providers (Gmail/Outlook/Yahoo), evaluate the reasons for ending up in Spam/Promotions based on the available signals in Customer.io and email headers. Create a priority correction plan.
2) Restructuring the structure of audiences and frequencies. Designing segments according to the life cycle (trial, onboarding, activated, dormant, churn risk, paid) with exclusion rules and frequency restrictions so that one contact does not simultaneously fall into conflicting campaigns. Accounting for time zones and working days (B2B).
3) Trigger chains. Collect/reassemble at least 3 automations in Customer.io based on events from Segment: onboarding (first 14 days), reactivation (no key event 10/21 days), lead nurturing for MQL (after application/demo). For each chain - a goal, entry/exit conditions, anti-duplicates, throttling, control metrics.
4) Experiments. Set up A-B testing of topics/preheaders and one content block (without full layout) with correct sampling and stopping criteria. Minimum of 2 parallel tests to obtain statistically valid results on volume.
5) Measurement and reporting. Set up a single set of events/conversions to evaluate the effectiveness of emails: activation in the product, request for a demo, upgrade, return to the product. Link campaigns/messages with UTM and agree on naming convention. Prepare a dashboard/report in the form of a document with before/after metrics and a plan for further iterations.

Specific result: corrected DNS settings and sending parameters, new segment scheme, at least 3 working trigger automations in Customer.io, 2 configured A-B tests, documented event map and naming, final report with recommendations for 30 days.

Measurable acceptance criteria (based on the results of 2 weeks after implementation, without taking into account seasonality): open rate for warm segments + 20% relative to the baseline, the share of delivered letters is not lower than 99%, a decrease in unsubscribes from mass mailings by at least 25%, a decrease in the share of spam reports (if available) by at least 20%, no duplicate sending between campaigns (checked on test contacts and on 100 random real profiles inside Customer.io). All changes are documented: what was changed, where, and how to roll back.

Important: the project requires confident work with Customer.io, Segment event model, deliverability practices (SPF/DKIM/DMARC, alignment, suppression, list hygiene) and understanding of B2B marketing lifecycle.