Definition

Email engagement strategy is the systematic approach to measuring, scoring, and acting on subscriber interaction levels to optimise deliverability and revenue. Subscribers who consistently open, click, and reply to emails tell mailbox providers that your messages are wanted. Subscribers who never engage tell providers your messages are unwelcome, lowering your overall sender reputation and causing deliverability problems for your entire list. An effective engagement strategy identifies which subscribers fall into each category and applies appropriate sending rules.

The standard engagement segmentation model defines five tiers. Active subscribers (20-30% of a typical healthy list) have opened or clicked within the last 30 days and receive full send frequency. Warming subscribers (10-15%) have shown recent first-time engagement after a period of inactivity — they are re-establishing a positive pattern. Passive subscribers (30-40%) receive emails but rarely engage — they receive reduced frequency to prevent fatigue without triggering disengagement. At-Risk subscribers (20-30%) have not engaged in 90-180 days and are moved to re-engagement workflows. Inactive subscribers (10-20%) have not engaged in 180+ days and are suppressed from regular sends entirely to protect deliverability.

RFM (Recency, Frequency, Monetary) models provide the analytical foundation for engagement scoring. Recency scores how recently the subscriber last engaged (days since last open or click). Frequency measures how often they engage within a defined window (e.g., number of opens in the last 90 days). Monetary measures the total value of purchases or conversions attributed to email engagement. Recency is the strongest predictor of future engagement — a subscriber who opened yesterday is far more likely to open again than one who opened six months ago, regardless of their historical frequency or monetary value.

Best Practices

Score engagement on a rolling 30-day window, not a fixed calendar period. A rolling window captures the subscriber's most recent behaviour patterns, which have the strongest predictive power for future engagement. Use a recency-weighted model where behaviour in the last 7 days is weighted 3x, days 8-30 weighted 2x, and days 31-90 weighted 1x. Older than 90 days receives zero weight for scoring purposes but remains in the subscriber history.

Suppress inactive subscribers (180+ days without engagement) from all regular promotional sends. Each inactive subscriber you mail increases your effective non-engagement rate. A 100,000-person list with 20,000 inactive subscribers suppresses your overall engagement rate by a critical margin, triggering mailbox provider filtering. Move these subscribers to a sunset re-engagement sequence with a maximum of 3 emails over 4-6 weeks, then permanently suppress non-responders.

Apply engagement tier-based send frequency limits. Active subscribers receive full send frequency (e.g., 12-16 emails per month). Passive subscribers receive reduced frequency (4-6 per month). At-Risk subscribers receive re-engagement series only (2-3 over 30 days). Inactive subscribers receive zero regular sends. These limits are expressed as maximums — if you send 15 emails per month to actives and only 5 to passives, you reduce complaint risk from the passive tier.

Trigger re-engagement programmes based on recency thresholds, not time-based calendars. A subscriber who has not opened in 60 days enters a win-back sequence, not after being subscribed for X months. Time-based triggers (e.g., "after 6 months of subscription") are less effective because they include subscribers who were actively engaging for 5 of those 6 months. The 60-day inactivity window resets after any open, click, reply, or forward.

Monitor engagement distribution trends over time. A healthy list maintains 40-50% combined active-plus-warming subscribers. If active share drops below 20% or inactive share rises above 25%, you have a list quality problem that requires immediate attention — likely caused by list acquisition practices that attract low-intent subscribers, insufficient sunsetting, or excessive send frequency driving fatigue.

What is a healthy engagement rate?

For active subscribers (30-day window), 25-30% is healthy. The overall list open rate (all tiers) typically ranges from 15-25% depending on industry. The metrics that matter most for deliverability are the engagement rate of the subscribers you actually mail and your spam complaint rate — keep both within healthy ranges.

How do I score engagement without purchase data?

Use recency and frequency of email engagement only — opens, clicks, and any positive signals (replies, forwards, add-to-address-book). Assign scores: +3 per open, +5 per click, +10 per reply, -5 per month of no engagement. This gives you a non-monetary engagement score that correlates strongly with future engagement and purchase likelihood.

When should I stop emailing an inactive subscriber?

After 180 days without any engagement (open, click, or reply), suppress the subscriber from regular sends entirely. Send a maximum of 3 re-engagement emails over 4-6 weeks. If no response after the third email, permanently suppress and move to a churn segment. Continuing to mail beyond this point damages sender reputation.

Does email engagement affect deliverability for transactional emails?

No, transactional emails (order confirmations, password resets, receipts) are not included in engagement-based filtering by most mailbox providers. You should not suppress subscribers from transactional sends even if they are inactive from marketing emails. However, you must ensure transactional emails remain genuinely transactional — including marketing content in transactional emails can cause them to be filtered as marketing messages.

What is the difference between engagement score and RFM score?

Engagement score is a broader measure that can include any positive or negative feedback signals. RFM score is a specific model using Recency, Frequency, and Monetary value. Engagement score is more appropriate for non-e-commerce programmes where monetary value is not tracked. Both models use recency as the most heavily weighted factor.