Engagement metrics beyond email opens: app logins, usage alerts and portal activity

Two residential accounts, same variable tariff, same town. One opened all twelve of last year's newsletters and switched away in March. The other opened none, logged into the app forty-one times, and is still on supply. Open rate ranked those two the wrong way round.

Email opens became the default engagement metric in utility marketing because they arrive free with the sending platform. The behaviour that predicts whether an account stays sits in four other places: the app, the alert stream, the self-service portal, and the permission layer that decides whether the first three reachreachThe number of unique people exposed to your message in a given period. Unlike impressions, reach counts each person once, no matter how often they see it.View full definition → anyone at all. This lesson defines that post-acquisition set and the benchmarks to argue with.

Why email opens mislead in this sector

Open rate measures curiosity, and since 2021 it barely measures that. Apple Mail Privacy Protection pre-fetches images for Apple Mail users, logging an "open" the human never made. On a list where most subscribers read in Apple Mail, the reported number is close to meaningless. Utility lists carry a second distortion: billing subject lines ("Your November statement is ready") get opened out of anxiety, so an account heading into arrears often looks more engaged than a contented one.

Logins, alert taps and portal sessions are self-initiated, repeated and time-stamped. They record that the customer came looking for you.

The metrics that matter

1. App login frequency

Definition: distinct login sessions per active app user over a period (usually monthly), segmented by cohort: new versus tenured, prepayment versus credit, solar versus non-solar.

How it's computed:

Monthly Active Usage Rate = (unique app logins in month / total registered app users) x 100
Login Frequency = total login sessions in month / unique users who logged in at least once

Benchmark (estimate, 2025 to 2026 range): leading US and European utility apps report monthly active usage in the 25% to 40% band among registered users, with power users (time-of-use tariffs, solar or battery assets) logging in 8 to 12 times a month. These are industry-reported estimates from digital transformation surveys (see Smart Energy Consumer Collaborative research), not regulatory figures, and they move with app maturity.

Worked example: 500,000 registered app users, 150,000 log in at least once in the month.

Monthly Active Usage Rate = (150,000 / 500,000) x 100 = 30%

If those 150,000 generate 900,000 sessions:

Login Frequency = 900,000 / 150,000 = 6 sessions per active user that month

Below 15% monthly active usage you have a discovery or usability problem: people registered once to pay a bill and never returned. Watch the denominator too. Registered users only grows, so a flat 30% can hide a shrinking active base behind a swelling registration count. Report against eligible accounts as well, and the honest number is usually several points lower.

2. Consumption alert click-through rateclick-through rateClick-Through Rate (CTR) is the percentage of people who click a link, ad, or call to action out of those who viewed it.View full definition → (CTR)

Alerts are push, SMS or email messages triggered by usage thresholds: "you have used 80% of your monthly budget", "unusual spike detected, check for a running appliance".

Alert CTR = (alerts clicked / alerts delivered) x 100

Benchmark (estimate): well-targeted usage alerts land in the 20% to 35% range, far above typical marketing email CTRs (often cited at 2% to 5%, per Mailchimp's benchmark data). The alert is triggered by the customer's own meter, not by a campaign calendar.

Opower built a business on exactly this. Its home energy reports compare a household against similar neighbours, and randomised evaluations of those programmes have repeatedly shown average consumption reductions of roughly 1% to 2%: small per household, large across millions. Oracle bought the company in 2016 and sells the alerting and reporting stack to utilities, so read vendor-published engagement figures with that in mind and insist on your own control group.

The commercial payoff is on the cost side. Customers who act on a high-bill alert are less likely to dispute the bill or call the contact centre, which pulls down cost-to-serve, the other half of the margin equation the acquisition lesson prices.

3. Self-service portal usage

Non-app web activity: bill viewing, autopay setup, meter reading submission, outage reporting, tariff changes.

Portal Adoption Rate = (customers who completed a self-service action in period / total customers) x 100

Benchmark (estimate): mature utilities often cite self-service adoption above 60% for routine tasks such as bill pay and autopay enrolment, per digital-channel reporting covered by Utility Dive. Complex actions (tariff switches, solar interconnection requests) stay far lower, often under 20%, because they need confidence as well as clicks.

4. Permission and notification opt-in

Every alert benchmark above is calculated on the reachable population, and reachability is a separate metric most teams never report. Track four things: push permission granted at install, permission still granted 90 days later, channel-level opt-out per campaign, and consent for granular meter data. In Great Britain, half-hourly readings need explicit customer consent, and the share who give it decides whether personalised alerts are possible at all.

A 30% alert CTR on a base where half the app users allow notifications reaches 15% of that base. Fix reach before optimising creative.

Tado shows where the ceiling comes from. Its customers open the app to change the temperaturetemperatureA setting that controls how random or predictable an AI model's output is: low keeps it safe and consistent, high makes it more varied and creative.View full definition →, so engagement is a by-product of the product doing its job. A supplier app with no control function should not borrow that cadence as a target. Ask instead what task your app owns that the customer cannot do anywhere else.

Why these predict retention better

Most residential customers cannot leave the wires or the pipepipeAll active sales opportunities across the stages of the sales process, together with their combined potential value and probability of closing.View full definition →. Churn is real only in competitive retail markets: Texas under ERCOT, Great Britain, Germany, parts of the Nordics and Australia. There, behaviour ranks accounts.

• Zero logins, no portal activity, notifications off: a dark account. You get no signal until the switch request arrives, and by then the only lever left is a retention offer.
• Regular logins and alert taps: a habit, and habits survive a competitor's price email.

So teams build a composite Digital Engagement Score across login frequency, alert CTR, portal adoption and reachability, then correlate it against 12 months of switch data before trusting it. Run that correlation first, because the score inverts in at least one cohort: prepayment customers log in constantly to top up, and that frequency tracks financial stress, not loyalty. Score them on the same scale as direct debit customers and the model will name your most fragile accounts as your best. Segment, then weight, then connect the result to the lifetime-value model the sibling lesson builds.

Building a simple engagement scorecard

Newsletter opens stay on the scorecard, heavily down-weighted. Email still carries messages; it should not be the primary engagement KPIKPIKey Performance Indicator, a measurable value that shows how effectively you're achieving a specific objective, tracked over time against a target.View full definition → (key performance indicator) reported to the board.

Then tier customers (high, medium, low) and run the retrospective: what was the actual switch rate in each tier over the last 12 months? That single analysis moves budget from subject-line testing to app and portal UX (user experience) faster than any argument.

🎬 [VIDEO: "How Utilities Use Data to Improve Customer Engagement" - youtube.com - search for utility customer experiencecustomer experienceThe overall perception a customer forms of your brand across every interaction, from first touch to post-purchase support.View full definition → case studies from industry channels like Smart Energy Consumer Collaborative or Utility Dive for real digital engagement examples]

A note on measurement discipline

AttributionAttributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → overlaps. A login two minutes after a high-bill alert is one behaviour, not two. Count the alert and the session it triggered as linked steps, or you will double-count your way into a flattering dashboard.

Seasonality does the same damage more quietly. Logins climb every winter in heating markets and every summer where cooling load dominates, so a January-to-December comparison flatters or condemns your app for reasons unrelated to any product work. Compare like months year on year and keep degree days on the same chart.

Key Takeaways

• Open rate is a weak retention signal here: inflated by Apple Mail Privacy Protection since 2021, and pushed up by billing anxiety on the accounts heading for trouble.
• App logins, alert CTR and portal adoption are self-initiated and time-stamped, which is why they rank accounts in switching markets (ERCOT, Great Britain, Germany, Australia).
• Estimated benchmarks: monthly active app usage 25% to 40%, alert CTR 20% to 35%, portal adoption above 60% for routine tasks and under 20% for complex ones. Confirm against your own platform data.
• Reachability is a metric, not an assumption. Push permission, its survival at 90 days and meter-data consent cap everything downstream.
• Validate a composite engagement score against 12 months of switch data before acting on it, and segment prepayment accounts out: their login frequency signals stress, not loyalty.
• Link alerts to the sessions they trigger, and read engagement trends against seasonal load rather than against last month.