How to audit a Klaviyo account: a 40-point checklist for ecommerce brands
Find broken journeys, missed revenue and avoidable cost. A 40-point Klaviyo audit across data, deliverability, flows, campaigns and measurement.
- #Klaviyo
- #Lifecycle
- #Checklist
Find broken journeys, missed revenue and avoidable cost. A 40-point Klaviyo audit across data, deliverability, flows, campaigns and measurement.

Frequently asked
Related resources
Solutions
Industries
See it in the product
The same loops this note describes run 24/7 against your customer base. Watch the workspace decide, experiment and execute 1:1.
Vocabulary
Definitions an answer engine can quote, each one a page of its own.
Compare
Neutral reads on the categories buyers evaluate against a decision layer.
All comparisonsTo audit a Klaviyo account, check data quality, channel consent, deliverability, list capture, segmentation, flows, campaigns, measurement and cost, in that order. A complete Klaviyo audit shows what is broken, which customers never get a useful message and which changes could add margin. The output is a prioritized list of actions, each with evidence, an owner and a way to validate it.
This guide is a 40-point Klaviyo audit checklist for ecommerce brands, grouped into ten areas. Not every point applies to every brand: a furniture store, a cosmetics brand and a food subscription need different journeys. Product details are linked to Klaviyo and Google documentation reviewed on 26 October 2026; the checks, scoring and prioritization are a recommended methodology, not official Klaviyo requirements or results from an audited account.
Log three kinds of findings separately:
For every check, record status, evidence, affected population, consequence, proposed action, dependency, owner and date. If you lack access to a piece of data, log the limitation instead of marking it correct. A good configuration is not the same as a profitable program, and not every available feature is an obligation.
Scope: use the last 90 days for recent operations, 12 months for seasonality and enough history to see your category's repurchase cycle. These are starting windows, not universal rules. Ask for read access to Klaviyo, the store, analytics, catalog and cost data, plus order and line-item exports. Run tests with labelled test profiles and orders; an audit does not require changing production sends.
Red flag: the goal is to raise the share of revenue attributed to email without knowing whether the business grows.
Red flag: a campaign bills more and leaves less contribution.
Red flag: new cohorts look worse only because they have had fewer days to buy.
Red flag: checkout abandonment shows heavy attribution, but nobody knows whether its discount creates additional orders.
Klaviyo detail: per the Shopify data reference, Placed Order includes tax and shipping and does not subtract cancellations and refunds the way Shopify does, and Refunded Order fires for full refunds only. Reconcile net revenue against the right source. [S1]
Red flag: purchase history and contact permissions belong to different identities.
Red flag: a cross-sell recommends a sold-out variant or an incompatible size.
Klaviyo detail: Shopify can send Placed Order and Ordered Product seconds apart. Mixing them between trigger and filters can cause unexpected skips. [S1]
Klaviyo detail: consent is stored per channel. A technically reachable contact is not the same as a contact with explicit consent. [S2] A configuration review is not a legal review.
Red flag: capture rate goes up while purchases or margin per visitor go down.
Red flag: the form with the most sign-ups is rewarded even though its leads do not buy.
Klaviyo detail: suppressed email profiles stop counting toward the active email profile limit, but other subscriptions can follow different criteria. [S3] Do not bulk-suppress only because someone does not open.
Reference: Gmail publishes different requirements by volume; bulk senders need SPF, DKIM, DMARC and unsubscribe conditions, among others. [S4]
Reference: Google recommends keeping the user-reported spam rate below 0.1% and avoiding 0.3%. Postmaster and Klaviyo denominators are not interchangeable. [S4]
Red flag: falling performance is offset by sending more to an ever less interested list.
Red flag: a polished preview hides a broken cart link or an empty block for part of the list.
Red flag: hundreds of segments exist, but all of them receive the same content and offer.
Red flag: flows work for those who enter, but leave many eligible customers outside.
Klaviyo detail: profile filters are checked at entry and before flow actions, and re-entry criteria live in the trigger settings. [S5] Removing skips indiscriminately can disable correct protections. [S6]
Klaviyo detail: Smart Sending has separate windows per channel; it is not a global email plus SMS plus push contact cap. [S7]
Red flag: a buyer receives 'make your first order' after ordering, or the reminder offers a better code.
Red flag: purchase intent that was already high gets an automatic discount.
Red flag: a review is requested before the order arrives, or the purchase is celebrated while a complaint is open.
Red flag: every category gets the same cross-sell on day seven, unrelated to product use.
Klaviyo detail: expected date of next order uses purchase behavior but not which products were bought. The documentation recommends specific flows when categories have different cycles. [S8]
Red flag: a customer is declared lost while still inside their normal repurchase cycle.
Mark these as not applicable if the business does not run such programs; never recommend them just to complete the checklist.
Klaviyo detail: the back in stock documentation warns that it cannot re-check item stock before a second native flow message. A reliable reminder may need extra logic. [S9]
Red flag: the calendar decides who gets what, regardless of customer state.
Red flag: clicks rise because the subject line promises something the landing page does not deliver.
Klaviyo detail: running out of uploaded codes can make a flow message skip, so auditing coupons is also auditing delivery. [S6]
Red flag: SMS looks profitable because it claims sales that would have come through email or direct purchase anyway.
Klaviyo detail: excluding Apple Privacy opens from attribution does not remove them from every open report, and attribution settings can recalculate history. [S10] Automated clicks are a different signal. [S11]
Klaviyo detail: the native Global Holdout feature requires, per the documentation reviewed, at least 400,000 total profiles. A custom control needs reliable assignment, consistent exclusions and independent analysis. [S12]
Klaviyo detail: a test can end on its date limit without reaching significance. The variant with the highest observed rate is not automatically a proven winner. [S13]
Red flag: an interesting correlation becomes a permanent automation without validation.
Klaviyo detail: reducing profiles does not guarantee an automatic plan downgrade; it depends on configuration and product. Advanced KDP counts profiles differently from the email plan. [S3]
Red flag: a former vendor still has send permissions, or an automation can change discounts without documented limits.
Red flag: a flow stops sending and nobody notices until the monthly report.
Deliverable: decisions for the next 30 days and a 60 to 90 day backlog. The audit is finished when you can act on it clearly.
This matrix forces you to review what is missing, not only what is configured. The interventions are candidates; whether they fit depends on data, consent, product and volume.
| State or signal | Candidate intervention | Exclusion or condition | Outcome to observe |
|---|---|---|---|
| Unidentified visitor | Contextual capture | Shopping experience and consent | Contribution per visitor |
| Subscriber without a purchase | Welcome and objection handling | Later purchase, sign-up source | First purchase per lead |
| Cart or checkout abandoned | Recovery and help | Purchase, checkout errors, pressure | Contribution per eligible profile |
| First order awaiting delivery | Information and expectations | Messages already sent by logistics | Issues and experience |
| First purchase completed | Complement or next product | Compatibility and usage cycle | Second purchase |
| Replenishment due | Relevant reminder | Quantity, subscription, recent purchase | Incremental repurchase |
| Requested variant sold out | Back in stock notice | Variant and sellable stock | Margin and satisfaction |
| Customer past their cycle | Contextual win-back | Seasonality and bad experience | Sustained reactivation |
| Issue or return | Resolution and trust recovery | Cause, support and return status | Resolution and retention |
| No sustained response | Lower pressure or sunset | Commercial cycle and external signals | List health and cost |
Do not give a missing alt text the same weight as an integration that duplicates orders. Use five statuses, and always show critical blockers and the share of verified checks next to any area score. A high average never compensates for a consent or data failure, and this editorial score is not a certified Klaviyo standard.
| Status | Meaning | Treatment |
|---|---|---|
| Correct | Enough evidence that it works | Keep and monitor |
| Improvable | Works; an opportunity or hypothesis exists | Estimate and prioritize |
| Failure | Breaks a necessary condition | Fix by severity |
| Not verifiable | Data, access or a test is missing | Resolve visibility |
| Not applicable | Does not fit the business model | Justify and exclude |
Before a test: eligible population × reachable exposure rate × assumed incremental effect × contribution per conversion, minus added costs. Every assumption needs a range and a source. If there is no basis for the effect, present the affected population and a test plan, not a euro figure.
After a randomized test: the difference in net revenue or contribution per assigned profile between treatment and control. Hypothetical example: 10,000 treated and 10,000 control profiles with average contribution of €4.80 and €4.50 give €0.30 per assigned profile, or €3,000 for the treated group before added costs. Report the uncertainty too; the averages alone do not prove a reliable effect.
The same holdout logic is explained in depth in our 90-day churn reduction plan, and the decision layer that ranks these actions is described on the retention decisioning and revenue discovery pages.
| Field | What to record |
|---|---|
| ID and area | Reference to the checklist point |
| Type | Failure, opportunity or hypothesis |
| Finding | What happens and to whom |
| Evidence | Export, configuration, test or sample, with period |
| Population and denominator | Who can be affected and how it was counted |
| Impact | Observed or estimated; never mix the two |
| Action and owner | Concrete change, scope, one accountable person and a date |
| Validation | Technical test or experiment, metric and guardrails |
Hypothetical finding: a cart recovery email uses a cart URL that drops the variant in one market. It reproduces with test orders, but the number of real sends affected is not yet counted. Action: fix the link builder, test variants and currencies, and review every message sharing the block. No recoverable revenue is claimed until reach and behavior are measured; the technical failure is verifiable even though its cost is not yet.
A complete Klaviyo audit delivers:
Markin today connects to Klaviyo and the store, surfaces coverage gaps and opportunities like the ones above, and prepares actions measured against a control. It does not automate all 40 points: consent, legal and several deliverability checks still need a human owner.
Official documentation reviewed on 26 October 2026. Confirm interface details and feature availability in your own account before turning this guide into click-by-click instructions.