Web push can reach a shopper without an open inbox or active browser tab. That immediacy makes the channel useful for restocks, price changes, time-sensitive delivery updates, and carefully selected promotions. It also makes fatigue easy to create.
What we see in lifecycle reporting is a channel reduced to opt-ins, deliveries, clicks, and attributed revenue. Ecommerce web push notification analytics should also ask whether permission was earned, demand was incremental, contribution remained positive, and the customer relationship survived.
Table of Contents
- Keyword decision and search intent
- Map the permission-to-profit funnel
- Build the web push scorecard
- Measure permission quality
- Prove incremental commercial value
- Control fatigue and frequency
- Technical and platform checks
- A four-week measurement plan
- EcomToolkit point of view
Keyword decision and search intent
- Primary keyword: ecommerce web push notification analytics
- Secondary intents: web push conversion rate, push opt-in quality, notification fatigue, incremental push revenue
- Search intent: evaluate and optimise a lifecycle channel
- Funnel stage: middle
- Page type: channel measurement framework
This framework treats browser permission as a scarce customer asset rather than an unlimited delivery list.
Map the permission-to-profit funnel
The channel begins before a message is sent:
- Customer reaches an appropriate permission moment.
- Permission prompt is shown.
- Customer allows, denies, or dismisses it.
- Subscription is created and remains valid.
- A campaign selects the customer.
- Push service accepts the message.
- Notification is displayed where measurable.
- Customer clicks, ignores, or changes permission.
- A commercial or service outcome occurs.
The browser Push API and Notifications API are separate building blocks: subscription and message delivery do not automatically prove that a notification was useful or produced a sale.
Build the web push scorecard
| Metric | Formula | Interpretation |
|---|---|---|
| Prompt acceptance | allowed ÷ permission prompts | timing and proposition |
| Subscription survival | active subscriptions after period ÷ created subscriptions | list quality |
| Provider acceptance | accepted sends ÷ attempted sends | technical delivery |
| Click-through rate | clicks ÷ measurable displays or accepted sends | message response |
| Conversion after click | purchases ÷ clicked sessions | post-click quality |
| Incremental contribution | holdout-adjusted margin less channel cost | business value |
| Permission loss | revoked or invalid subscriptions ÷ active base | fatigue/decay |
| Complaints or support contacts | channel-related contacts ÷ recipients | customer cost |
Be explicit when “display” is unavailable or inconsistent. Use accepted sends as a denominator only if the dashboard labels it honestly.
Measure permission quality
An immediate prompt on the first page can produce accidental denials and low-context opt-ins. Ask after a relevant action: requesting a restock alert, tracking an order, saving an item, or choosing to receive offers.
Compare permission cohorts by:
- prompt context;
- pages viewed before prompt;
- new versus returning visitor;
- logged-in versus anonymous state;
- device and browser;
- requested notification purpose;
- market and language;
- subscription survival after 7, 30, and 90 days.
Do not optimise acceptance without downstream quality. A smaller restock cohort with clear intent may produce more value and less permission loss than a large generic promotional list.
Prove incremental commercial value
Last-click revenue can over-credit push. A customer may already have intended to buy, received email and push for the same campaign, or returned through another channel.
Use recipient-level holdouts. For eligible customers, suppress the push for a stable random group while leaving other treatment rules unchanged. Compare:
- purchase rate and net revenue;
- contribution margin;
- time to purchase;
- promoted-item and substitute-item demand;
- cancellations and returns;
- email/SMS response;
- permission loss;
- repeat purchase over a longer window.
Google Analytics ecommerce events can capture product and purchase outcomes, but experiment assignment and send evidence usually belong in the warehouse. Keep stable item and transaction identities using Google’s ecommerce measurement guidance.
| Result | Clicks | Incremental contribution | Decision |
|---|---|---|---|
| High | High | Positive | scale within frequency limits |
| High | Low | Positive | improve message-to-landing match |
| Low | High | Weak/negative | reduce incentive or targeting waste |
| Low | Low | Negative | stop or redesign |
Control fatigue and frequency
Frequency is a customer-level property, not a campaign setting. A person can qualify for restock, browse, cart, price-drop, editorial, and promotional messages from different automations.
Create a central pressure ledger:
| Field | Purpose |
|---|---|
| customer/subscription key | total contact pressure |
| message purpose | service, triggered, promotional |
| send and display time | recency and clustering |
| campaign priority | conflict resolution |
| click and conversion | observed response |
| permission state change | channel damage |
| other-channel contacts | cross-channel pressure |
Set limits by purpose and engagement. A delivery exception is not equivalent to a generic promotion. Suppress low-priority sends when a higher-priority service or trigger message is active.
Use our back-in-stock analytics guide to connect notifications to inventory truth, and our email deliverability framework for a complementary owned-channel view.
Technical and platform checks
Evaluate:
- permission prompting control;
- service-worker ownership and update behaviour;
- subscription expiry and endpoint cleanup;
- identity stitching after login;
- campaign deduplication and idempotency;
- market, timezone and quiet-hour controls;
- deep-link and fallback behaviour;
- consent and preference separation by purpose;
- export access for sends, clicks, failures and costs;
- deletion propagation when a customer withdraws or requests erasure.
Treat endpoints and customer mappings as sensitive data. Limit access, avoid exposing raw subscription details in broad dashboards, and document retention.
A four-week measurement plan
Week 1: map permission, subscription, send, provider, click, purchase, refund and permission-loss events.
Week 2: validate denominators and reconcile vendor exports with first-party events. Segment permission cohorts by acquisition context.
Week 3: introduce recipient holdouts and a cross-channel pressure ledger. Report contribution rather than attributed revenue.
Week 4: adjust prompt timing, frequency, quiet hours and automation conflicts. Retire campaigns that create clicks without incremental value.
Review outcomes over enough time to observe permission decay and returns. Immediate response is only one part of channel health.
EcomToolkit point of view
Web push is valuable when it delivers a relevant reason to return at the right moment. It is destructive when permission becomes a shortcut around customer attention.
Optimise for incremental profit per permission while protecting the future usefulness of the channel. Claim a free EcomToolkit audit to connect push permissions, campaign evidence, ecommerce outcomes, and cross-channel fatigue.
