How To Segment By Opens And Clicks

Open and click data are among the most accessible signals in an email platform. They can help a sender decide who should receive a new launch, a higher-frequency editorial stream or a re-engagement message. Used carelessly, they can also create misleading audiences: people who appear highly engaged because their mail client fetched a tracking pixel, or people who look inactive because they read emails without loading images.

The practical answer is to treat opens and clicks as signals, rather than proof that a person read, wanted or acted on an email. Clicks generally provide a stronger indication of deliberate interest when they lead to meaningful content, but they need validation too. A sound segmentation model combines message activity with customer, website and consent data, then changes its rules to fit the sending cadence and buying cycle.

What open and click segments actually measure

An open is normally recorded when an email client requests a tiny remote image, often called a tracking pixel. It does not prove that the recipient read the message, understood it or even saw it in the inbox. Image blocking, text-only reading and some privacy settings can hide genuine reading. Conversely, privacy features can create an open event without the recipient viewing the message in the conventional sense.

Apple’s Mail Privacy Protection is the clearest reason to avoid treating opens as a literal reading measure. Apple says the feature privately downloads remote content in the background and prevents senders seeing whether a message has been opened; it also hides the recipient’s IP address. It does not apply to links, so click data is affected differently. See Apple’s explanation of Protect Mail Activity and its Mail Privacy Protection guidance for iPhone.

A click is recorded when a recipient follows a tracked link, typically through a redirect controlled by the sending platform. This usually indicates more intent than an open, particularly when the destination is a product page, booking journey, article or account area. However, a click may be generated by a corporate security tool checking links, a prefetching process, an accidental tap, or a recipient who is simply looking for an unsubscribe or preference option.

Neither metric is a direct measure of inbox placement. Google explicitly says it does not track opens and cannot verify third-party open-rate accuracy; low open rates alone are not a reliable sign of a deliverability problem. Gmail’s sender guidance instead directs senders towards recipient feedback, authentication, reputation and spam-rate monitoring. Google’s email sender guidelines are a useful reminder that engagement reporting should sit alongside operational delivery data, rather than replace it.

Start with clean event definitions

Before building an audience, document exactly what each event means in your platform. Different systems may record a unique open once per campaign, count several opens, attribute clicks differently after forwarding, or retain historic events after a contact has changed status.

For most permission-based marketing programmes, distinguish at least the following:

Also decide which email types may contribute to engagement. A password-reset click, delivery tracking click or receipt open should not automatically qualify a person for more promotional mail. Transactional messages solve a different problem and may be read even by someone who no longer wants marketing.

Use recency and intent, not just totals

A segment labelled “opened five emails” is rarely useful on its own. Five opens in the past week suggest something different from five opens spread over two years. Segment rules should therefore include a time window and, where possible, the type of interaction.

The appropriate window depends on how often you send and how long customers take to decide. A daily publisher might assess recent engagement over 14 to 30 days. A retailer sending weekly may use 60 to 120 days. A business with quarterly communications or a long consideration period may need a much longer view. The point is not to copy a universal 30-, 60- or 90-day threshold; it is to give a willing subscriber enough realistic chances to engage before reducing contact.

Intent matters as well. A click on “view availability”, “continue application”, “read the full guide” or a category link is usually more informative than an image click, social icon or footer link. Keep preference-centre, privacy-policy, customer-service and unsubscribe links out of a positive-engagement definition unless there is a specific analytical reason to retain them separately.

A practical segmentation framework

Rather than maintaining dozens of fragile micro-segments, begin with a small number of groups that lead to a clear sending decision. The exact names do not matter; the rules and treatment do.

Recently click-engaged

This group includes people with one or more qualifying marketing clicks in a recent, cadence-appropriate window. If an on-site event is available, require or reward a visit beyond the redirect. It is commonly the safest group for an important campaign, an increase in frequency or the first stage of a controlled volume ramp.

Do not assume every click is human. Where your tools provide the evidence, exclude clear automated patterns: for example, numerous different links requested within seconds, a recognised security-scanner user agent or IP range, and no subsequent browser or site activity. Avoid a crude rule that removes every first click, as it may discard real subscribers.

Recently open-engaged but not click-engaged

This is a useful audience for content-led messages, newsletters and low-pressure reminders, but it should not be treated as identical to the click-engaged group. Open-only engagement may represent real readership, especially for editorial content whose value is largely contained in the email. It may also include privacy-generated events.

Use this segment to test subject lines, format and relevance. For promotional sends, keep the volume and frequency proportionate, and watch unsubscribe and complaint signals rather than assuming the open rate validates the programme.

New subscribers and recent customers

Newly opted-in contacts need a grace period. They have not had enough opportunity to generate a meaningful open or click history, and suppressing them as “inactive” can undermine a welcome programme before it starts. Keep their consent source, sign-up date, declared interests and first-party purchase or account activity available as separate fields.

Likewise, a recent purchaser may be valuable even if they have not clicked a recent campaign. Suppress them from inappropriate acquisition offers, but do not move them to a low-engagement segment simply because they have been busy using the product rather than opening email.

Fading engagement

This group has older opens or clicks but no recent qualifying activity. It is often the right place for a lower-frequency content stream, a preference prompt or a carefully scoped win-back test. A fading segment should be smaller and less frequent than the active group; repeatedly sending the same full volume to people who have stopped responding is rarely a route to better results.

Give a re-engagement email a genuine purpose. Ask subscribers what they want to receive, invite them to update interests, or offer a concise selection of useful content. “We miss you” alone does not establish relevance.

Inactive or unconfirmed engagement

These contacts have no qualifying click, no dependable open activity and no recent first-party action over a period appropriate to the programme. They should normally be excluded from routine promotional mail, at least temporarily. A limited re-permission or re-engagement approach may be suitable for contacts who originally gave valid permission, but it should have an end point. If there is still no response, suppressing rather than continuing to mail protects both audience quality and sender reputation.

Google’s Postmaster guidance recommends sending only to recipients who want the messages, periodically confirming continued interest, and considering the removal of people who do not open or read email. It also warns that a low reported spam rate can coexist with poor inbox placement when much mail is already being routed to spam. Review Google’s Postmaster Tools dashboard guidance when interpreting engagement alongside reputation and complaint data.

Build the logic in the right order

Segment conditions are easier to manage when they follow an order of precedence. Start with exclusions that must always win: global unsubscribes, hard bounces, complaints, legal or contractual suppression rules, and contacts without the required marketing permission. Apply contact-pressure limits next, so a person cannot be selected for several overlapping campaigns in the same period.

Then assess lifecycle state, such as new subscriber, active customer or lapsed customer. Only after that should you apply engagement rules. This prevents a recent transaction or account action being erased by a simplistic “no opens in 90 days” filter.

A platform-neutral version of the central rule could look like this:

1. Contact has permission for the proposed message and is not suppressed.
2. Contact has not received the maximum number of marketing messages permitted in the pressure-control window.
3. Contact has clicked a qualifying marketing link in the previous defined period or completed a recognised first-party action.
4. If neither is present, include a smaller open-only group only where the campaign is suitable and the send frequency is controlled.

Keep the date field explicit: “most recent qualifying click date” is more reliable than a generic “engaged” flag whose meaning has changed over time. Where possible, preserve the campaign ID, link category and source of the event too. This makes it possible to audit why a contact was selected and to update rules without rebuilding history.

Do not use opens and clicks in isolation

Good segmentation joins behavioural events to information that explains them. Useful fields may include stated topic preferences, product category, location where relevant and permitted, customer status, purchase recency, account usage, language, sign-up source and the email stream the person chose. An open or click becomes much more useful when it answers a specific question: which content does this subscriber prefer, which stage are they at, and how often are they likely to welcome messages?

For example, a subscriber who clicked several beginner guides may belong in an educational journey rather than a generic “highly engaged” segment. A customer who clicks support content after purchase may need onboarding help, not another discount. Segmenting only by the volume of activity can flatten these differences.

Separate marketing streams by purpose as well. Someone who actively reads a monthly insight newsletter may not want frequent product promotion. Separate suppression and frequency rules allow a sender to respect that distinction.

Measure whether the segments improve the programme

After deploying a new model, compare groups using outcomes that reflect the campaign’s purpose: qualified visits, conversion, revenue where attribution is sound, account activation, preference updates, unsubscribes and complaints. Compare like with like: the same campaign family, audience type and time period.

Pay particular attention to what happens when a segment is expanded. Adding open-only contacts may increase the apparent audience and open volume, but it can also dilute click-through performance or raise unsubscribes. That is not necessarily a reason to abandon the group; it may mean the content, frequency or offer needs to differ.

Monitor deliverability at provider level where data is available. Gmail recommends maintaining user-reported spam rates below 0.1% and avoiding 0.3% or higher, while also requiring authentication and easy unsubscribing for larger senders. Its sender-guidelines FAQ explains that these figures are updated daily and that a high rate can affect delivery. Engagement segments should help lower unwanted-mail signals, but they cannot compensate for weak consent, broken authentication, poor list hygiene or an excessive sending cadence.

When engagement data needs a deliverability review

Unexpected engagement changes should be investigated before audience rules are rewritten. A sudden shift in reported opens can reflect a change in audience mail-client mix, privacy settings or tracking configuration. A sharp fall in clicks may be a content issue, a broken redirect, poor inbox placement, a slow landing page or a mismatch between the promise in the subject line and the message itself.

Where a sender has a reputation or inbox-placement problem, merely narrowing to recent openers can hide the symptom rather than resolve the cause. The underlying review should cover permission and acquisition routes, suppression handling, authentication, sending infrastructure, provider feedback, complaint patterns and frequency. Email Foundry’s work is relevant at this point because audience verification, likely-human engagement filtering, contact-pressure controls and provider-level delivery monitoring can help distinguish weak audience data from a wider deliverability problem.

The goal is not a perfect definition of engagement; email data cannot provide one. The goal is a transparent, conservative model that rewards recent, meaningful signals, gives new and long-cycle subscribers fair treatment, and stops routine marketing to people who no longer show evidence that they want it.