The authentication baseline every Shopify store needs
Send from a dedicated email subdomain (e.g. mail.yourbrand.com) so marketing reputation stays isolated from your corporate domain. Publish SPF and DKIM for that subdomain, align them with a DMARC record, and verify inside the ESP before any large send. Shopify’s own notification domains handle transactional mail, so the marketing subdomain is the one you control, warm, and monitor. Skipping DMARC does not break the inbox directly, but it removes the feedback loop that tells you spoofing or misalignment is happening.
Warming is behavioral, not calendar-based: new subdomains earn volume by consistent audience engagement, not by waiting. Start with flows to engaged recent subscribers (welcome, cart, post-purchase), hold broadcasts until two weeks of stable complaint and bounce data exist, and never ramp volume into a peak season on an unproven domain.
The three failure modes behind most placement drops
Placement complaints cluster into three diagnosable families. Audience state: the list widened faster than engagement — popup sweeps, purchased-sounding imports, sunset segments never built. Infrastructure: authentication changed, alignment broke, or new infrastructure was never warmed. Content and cadence: the same offer repeated until recipients said “spam”, or message frequency crossed the tolerance of everyone including the engaged cohort. Three families, three first checks — which is why cordoning the symptom before buying a tool matters more than the tool itself.
List hygiene as a margin decision
Every billing model punishes list bloat somewhere: active profiles on Klaviyo, stored contacts on Mailchimp and Omnisend, people-in-account on Drip. Suppressed and unengaged contacts either bill or drag placement — sometimes both. A weekly hygiene rhythm costs thirty minutes: review the engaged ratio, sunset the stale cohort, log the complaint rate, and reconcile Shopify unsubscribes against the ESP. Stores that do this weekly pay less per delivered email than stores that clean before renewal negotiations.
Purchaser suppression participates in hygiene too. Recent buyers receiving cart or browse messages are a margin leak and a complaint generator in one — the flow that “works” while discounting completed checkouts is the worst kind of false positive. Verify purchase and refund exits with staff profiles whenever the catalog, pixel, or fulfillment integration changes.
Decision table: symptom to first control
| Symptom | Most likely control gap | First check | Tool-dependent fix |
| Falling Gmail placement | Reputation or authentication | DMARC reports, SPF/DKIM alignment | Dedicated IP or subdomain discipline |
| Complaint rate above 0.3% | Audience-state control | Which flow/campaign sent most complaints | Sunset flows, suppression rules |
| Hard bounces rising | Capture or import hygiene | Signup sources and recent imports | Double opt-in, validation on import |
| Opens fine, clicks collapsing | Content or targeting drift | Segment freshness, offer repetition | Engagement-gated sending |
| SMS opt-outs spiking | Channel consent boundaries | Frequency caps and quiet hours | Channel-specific suppression |
| Deferrals at one inbox provider only | Domain reputation with that provider | Postmaster tools, volume ramp history | Throttled sends, subdomain split |
| Placement fine, revenue flat | Not a deliverability problem | Offer, segment, margin panel | Stop tool-shopping; fix the offer |
| Complaints cluster on one campaign type | Content/consent mismatch | Who received it and why | Engagement-gated targeting |
Pricing note: what deliverability capability costs
Deliverability features ride inside ESP plans: Klaviyo includes subdomain authentication and monitoring; Brevo gates dedicated IPs by tier; Mailchimp’s audience model makes hygiene a financial decision every month. Because deliverability tooling is rarely a separate line item, the practical cost question is operator time — roughly one structured hour weekly per 10k active recipients in smaller stores. Check each vendor’s official pricing page for dedicated IP, deliverability services, and monitoring add-ons before assuming they are included; plan gates change and quotes should be recorded with dates.
Peak-season deliverability runbook
Sixty days out: authenticate (re-verify), sunset unengaged, confirm purchase/refund suppression with test orders, and baseline complaint and bounce rates per flow. Thirty days out: warm any new subdomain, rehearse the highest-volume campaign to an engaged-only segment, and confirm SMS quiet-hour and frequency rules if texting. During the peak: lock flow architecture (no new branches mid-season), watch deferrals by inbox provider, and keep one person on rotation for complaint spikes. After: reconcile unsubscribes, archive the incident notes, and carry the baseline into next season. Stores that skip the sixty-day step are the ones making DNS changes in November — the most expensive possible moment.
Weekly operating rhythm: the thirty-minute version
Monday, ten minutes: export the week’s complaint and bounce rates by flow; anything above your written threshold becomes a named task. Ten minutes more: review the engaged ratio and age the unengaged cohort toward sunset. Five minutes: reconcile Shopify unsubscribes and refunds against ESP states — every mismatch is a suppression bug. Five minutes: check that no shopper received more than your per-day message cap across flows (most tools can query this; the one that cannot is telling you something). Leave the last five minutes for notes: what changed, what to watch. Deliverability problems are weekly-shaped; monthly reviews find them a month late. Write the thresholds down once — complaint rate 0.2% review, 0.3% action, bounce 2% investigate — so the review doesn't depend on whoever remembers last quarter's panic.
Tool-team fit table
| Team shape | Realistic deliverability posture | Tools that fit it | Avoid |
| Founder-operated, under 5k list | Simple subdomain, weekly complaint check, conservative volume | Shopify Email, Omnisend, tinyEmail | Elaborate infrastructure nobody maintains |
| One marketer, 5k–25k list | Automated sunset segments, monthly hygiene hour, DMARC monitoring | Klaviyo, Sendlane, Brevo | Full-list imports with consent skipped |
| Retention team plus agency | Documented runbook, incident thresholds, quarterly re-certification | Klaviyo, Attentive, ActiveCampaign | Undocumented agency-owned suppression |
| SMS-heavy programs | Channel-separated consent, carrier-fee aware caps | Postscript, Yotpo Email & SMS | Treating email permission as text permission |
Two field scenarios
A $40k/mo beauty store saw Gmail placement degrade two months after a popup sweep doubled its list. The tools panel said nothing was “wrong” — authentication was fine. The engaged ratio had fallen from 51% to 29%, complaint rate crossed 0.25%, and the sunset flow had never been built. Two weeks of hygiene: sunset 9k dormant profiles, rebuild engagement-gated cohorts, cap per-shopper messages at one per day across flows. Placement recovered without touching the vendor; the fix was operating rhythm.
A $75k/mo fashion brand migrated ESPs and immediately saw deferrals — classic unwarmed-new-domain behavior. They had skipped the warm-up because “the old domain was fine”. The recovery was a two-week ramp on flowed email to engaged segments before any broadcast, plus explicit DMARC verification. The migration guide exists to prevent this exact sequencing error: authentication and warm-up are part of the migration, not an afterthought nobody planned.
Measuring deliverability without fooling yourself
Open rates are unreliable with Apple Mail privacy protection inflating machine opens. Build the real panel from five numbers: delivered rate (not accepted — delivered after deferrals), complaint rate by domain, hard bounce rate, click-to-delivered rate on a fixed engaged segment, and revenue per delivered email. Track them weekly in one spreadsheet a non-marketer could audit. When platforms claim “99% inbox placement”, ask which seed test produced it and when — seed panels correlate poorly with recipient-domain placement.
Margin closes the loop. A recovery flow with 96% delivered and a 15% discount on every recovered order may be worse for contribution than a 92% delivered flow with reminders only. Deliverability reviews that stop at placement treat symptoms of a business problem; the complaint rate is as much a function of what you say as of where it lands.
Internal checks and official sources
Use the consent checklist, migration playbook, flow-collision guide, and pricing-model guide alongside the official links in each profile. Record the access date, plan, limits, add-ons, data retention, and cancellation terms; do not preserve a stale price claim in a procurement document. For the sending-risk side of the same coin, see the deliverability use case and the Black Friday preparedness playbook. For pairwise tool tests, start with Klaviyo vs Mailchimp and Brevo vs Mailchimp.