The Self-Healing Pixel
Mass ships a first-party tracking pixel on every published funnel, then closes the loop: it watches the metrics that pixel produces against a set of healing rules and — depending on how you set it — alerts you, proposes fixes, or auto-generates new copy and assets when performance slips.
12 min read · Updated for the current release
What the Self-Healing Pixel is
Two systems working as one: a first-party analytics pixel that captures behavior, and a rules engine that acts on it.
Most analytics stop at a dashboard — they tell you something is wrong but leave the fix to you. The Self-Healing Pixel is built to act. The pixel collects a clean, first-party stream of visitor events from your live funnel, and the healing engine continuously evaluates those events against rules you control, taking the action you've authorized when a rule trips.
Because Mass also generated your funnel, it can do something a generic tracker can't: when conversion drops, it already knows your offer, audience, and brand voice, so it can write genuinely on-brand alternative headlines, CTAs, subject lines, and ad copy as the remedy.
- First-party — events post to your own Mass endpoint and land in your tenant's data — no third-party dependency required.
- Always-on — the pixel fires on every published funnel page automatically once tracking is set up.
- Actionable — rules don't just flag problems; they can trigger AI generation of fixes scoped to your offer.
The tracking pixel
A lightweight script on every published page that records what visitors actually do.
When a funnel goes live, Mass injects a tracking script that captures the visitor journey and posts events to the analytics endpoint, where they're stored per-tenant for the campaign dashboard, the CRM activity feed, and the automation triggers to read. It identifies an anonymous visitor and session, dedupes the initial page view per tab, and captures engagement as the visit unfolds.
- Page views — fired once per page load per tab, with page URL, title, path, and referrer.
- Scroll depth — the deepest scroll percentage the visitor reaches, sampled as they move down the page.
- Clicks & form submits — interaction events so you can see what's being engaged and where conversions happen.
- Time on page & exit — a beacon on unload records total time on page and final scroll, the one reliable way to capture exits.
- Attribution — UTM source/medium/campaign/content/term plus device, browser, and OS travel with every event.
Conversions are first-class
Form submits and conversion events are captured alongside engagement, so the healing engine can reason about conversion rate, not just traffic — and high-intent activity flows into the CRM as buying signals.
Healing rules
Each rule pairs a condition on a live metric with the action to take when that condition is met.
The healing engine ships with a set of sensible defaults you can toggle on or off per campaign. A rule watches one metric, defines the threshold that counts as a problem, and names the remedy. When you run a check (or the monitor runs one on schedule), every enabled rule is evaluated against the campaign's current KPIs and its result is written to an activity log.
- Low Conversion Rate — when conversion drops below 1%, generate alternative headlines and CTAs.
- High Bounce Rate — when bounce exceeds 70%, analyze page speed and suggest content improvements.
- Low Email Open Rate — when opens fall below 15%, generate new subject-line variants.
- High Cost Per Lead — when CPL runs 50% over target, suggest ad targeting and copy refinements.
- Low Scroll Depth — when average scroll is under 40%, suggest reordering the page for engagement.
- Stale Content — when there are no conversions for 72 hours, flag for a refresh and notify the owner.
Monitoring & automation modes
You decide how hands-on the system is — from a quiet alert to fully automatic fixes.
Turn on auto-monitoring and pick a cadence — hourly, every six hours, daily, or weekly — and the engine runs your enabled rules on that schedule. The critical choice is the action mode, which controls what happens the moment a rule trips.
- Notify only — log the issue and alert you. Nothing changes until you act — the safest setting.
- Suggest fixes — the engine drafts recommended changes (new headlines, CTAs, subject lines) for you to review and apply.
- Auto-apply — for failed rules, Mass calls the generate-alternatives endpoint and produces on-brand replacement assets automatically.
Health score & activity log
A status overview shows the campaign health score, the number of active rules, and how many issues were detected. Every check appends to an activity log so you have a running history of what tripped, when, and what was done about it.
Setting it up
Self-healing rides on tracking, so the order is: add the pixel, go live, then enable healing.
- 1
Set up tracking
On the Fast Launch checklist, complete Setup Tracking — add your pixel/analytics IDs so the funnel emits events.
- 2
Go live
Publish the funnel. The pixel starts collecting visitor events and the campaign dashboard begins populating KPIs.
- 3
Open Self-Healing
In the campaign dashboard, open the Self-Healing panel, review the default rules, and toggle the ones you want.
- 4
Choose a cadence & mode
Enable auto-monitor, pick a frequency, and set the action mode (notify, suggest, or auto-apply).
Self-healing for AEO content
The same closed-loop idea extends to your AI-citable blog posts, with conservative guardrails.
Beyond funnels, Mass runs a separate self-healing and self-expanding loop for AEO content. A daily tick reads a dual-signal feed for each published post — AI-answer visibility and Google rank — scores whether the post is winning or losing on each, and enqueues a heal or expand job inside a per-project weekly budget.
The guardrails are deliberately cautious: posts are never healed in their first 30 days, require a meaningful history of rank data before any action, observe a cooldown between attempts, and honor an operator's rejection for a set period. The cron only scores and queues — the actual rewriting happens out-of-band so the loop stays cheap and predictable.
Two loops, one philosophy
Whether it's a funnel KPI or a post's citation share, the pattern is the same: measure continuously, compare against thresholds, and act with the level of autonomy you've authorized.