Look beyond the campaign calendar
An email marketing agency should explain which customer behaviours it intends to influence, what data makes that possible, and how it will judge success. A polished campaign calendar is a production plan, not a retention strategy.
When comparing email and retention agencies, start with the commercial problem. Improving product adoption, encouraging another purchase, and preventing avoidable cancellations require different programmes. The proposal should connect that problem to specific audiences, messages, exclusions, and outcomes.
Do not reward a long automation list by itself. Ask why each journey belongs in your business and what would cause the agency to change or retire it.
Define lifecycle scope before comparing proposals
Map the customer journey and its exceptions
Ask each shortlisted agency to describe the lifecycle it would manage. Depending on the business, that might include:
- Permission-based acquisition and welcome messages.
- Product education, onboarding, and activation.
- Browse or checkout recovery where appropriate.
- Post-purchase guidance and service handoffs.
- Replenishment, renewal, and repeat-purchase journeys.
- Lapsed-customer reactivation and suppression.
The exceptions matter as much as the happy path. What happens after a refund, an unresolved support complaint, a cancellation, or a purchase through another channel? A credible proposal explains how those events change eligibility and messaging.
For retail buyers, our guide to choosing an ecommerce marketing agency helps frame the wider relationship between retention, acquisition, and the store experience.
Separate strategy from implementation
Require explicit ownership of data integration, event definitions, copy, design, template development, quality assurance, deployment, and reporting. Establish who fixes broken tracking and who approves changes to customer eligibility.
Ask whether the engagement includes ongoing journey maintenance or only initial builds. A replenishment sequence can become inappropriate when stock availability changes; an onboarding sequence needs revision when the product changes.
The agency should also identify dependencies outside its control. If weak product education or a confusing checkout prevents conversion, more email is not necessarily the answer. Coordination with conversion optimization agencies may be more useful than increasing send frequency.
Treat deliverability as an operating discipline
Deliverability deserves a named owner, an escalation process, and access to relevant diagnostics. Ask who checks authentication, monitors mailbox-provider feedback, investigates failures, and has authority to pause sending.
Google's email sender guidelines distinguish requirements applying to all senders from additional requirements for bulk senders, including authentication and unsubscribe provisions. A strong agency should explain which requirements apply to your programme rather than offer a generic assurance that the platform handles everything.
Useful deliverability questions include:
- How are acquisition sources, permission records, and suppression rules audited?
- How are hard bounces, complaints, and unsubscribes handled?
- How are promotional and transactional message streams managed?
- What changes when sending volume or audience composition changes?
- Which signals trigger investigation or a pause?
Ask for a redacted incident report. Look for a clear diagnosis, action taken, customer impact, and prevention plan—not a promise that inbox placement can be guaranteed.
Read published guardrails correctly
Mailbox-provider guidance supplies useful boundaries, but these are not agency performance benchmarks:
- **Gmail's recommended operating level:** Google advises keeping spam rates reported in Postmaster Tools below **0.1%**, according to its sender guidelines.
- **Gmail's upper boundary to avoid:** Google says senders should avoid ever reaching a spam rate of **0.3% or higher**, in the same sender guidelines.
- **Yahoo's bulk-sender requirement:** Yahoo specifies keeping spam complaint rates below **0.3%** in its sender requirements and recommendations.
These are distinct published guardrails, not independent observations of campaign performance. Ask the agency to explain each provider's reporting definitions; do not average the values into a universal target. Staying below a boundary does not establish that messages are relevant or commercially effective.
Ask for metrics that reveal judgement
Business outcomes before dashboard totals
The reporting plan should begin with the business model. For commerce, ask about repeat purchasing, customer-cohort revenue, and contribution after discounts and relevant costs. For subscription businesses, ask about activation, renewal, cancellation, and sustained product use.
Require definitions alongside results. Which customers enter a cohort? What counts as a repeat purchase? Are refunds included? Has the observation period changed? Without stable definitions, a better-looking dashboard can conceal a different calculation.
Our guide to what ecommerce brands should expect from a retention marketing agency provides additional context for setting the remit.
Separate attribution from incrementality
Platform-attributed revenue answers a credit-assignment question under the platform's rules. Incrementality asks what happened because of the marketing that would not otherwise have happened.
Ask candidates to explain their attribution window, treatment of overlapping channels, and approach to customers already likely to buy. Where feasible, request randomized holdouts or another defensible comparison, with exclusions and limitations agreed before launch.
A capable agency should acknowledge when the available audience or purchase cycle makes a reliable experiment difficult. It should not relabel attributed revenue as proven incremental revenue.
Use engagement as diagnostic evidence
Apple explains that Mail Privacy Protection can download remote content in the background regardless of whether the recipient engages with the message, in its Mail Privacy Protection notice. That makes open-based reporting an unreliable standalone measure of human attention.
Ask how the agency uses downstream behaviour to interpret engagement and checks whether automated activity affects click reporting. Campaign comparisons should account for audience selection, offer changes, seasonality, and send frequency rather than crediting the creative automatically.
Test the operating process before signing
Request redacted work samples that show reasoning, not just visual polish:
- A lifecycle map with entry conditions, exits, and exclusions.
- A test brief stating the hypothesis and decision rule.
- A quality-assurance checklist covering links, rendering, personalization, and eligibility.
- A report distinguishing observed outcomes from causal claims.
- A handover plan for templates, documentation, accounts, and customer data.
Ask who approves sends and who owns suppression decisions. The US Federal Trade Commission's CAN-SPAM compliance guide explains that businesses cannot contract away their legal responsibility simply by hiring another company to handle email marketing. Treat jurisdiction-specific compliance as an explicit responsibility, with legal advice where needed.
Buy a documented operating model
The contract should separate initial auditing and implementation from recurring production, optimization, and incident support. Clarify software charges, migration work, approval delays, and what happens when integrations fail. Compare the arrangement against the trade-offs in agency pricing models, rather than assuming every retainer covers the same work.
Choose the agency that can explain its decisions, expose its measurement limitations, and leave you with usable documentation. Good retention work is not demonstrated by sending more messages. It is demonstrated by a controlled programme that learns which messages genuinely help customers and the business.