Signal or Noise: Tearing Down Zomato Session Recording Notes

Marketing Academy · Field Work●Conversion Rate Optimization

MiniTeardown· 20 minutes

Zomato

Objective: Given 5 anonymized session-recording notes from a food-delivery checkout flow, distinguish real friction signals (rage clicks, hesitation, scroll-backs) from normal browsing behavior, per the lesson's Step 3 framework.

You're reviewing a batch of Hotjar session recordings for Zomato's restaurant-checkout flow after a spike in cart abandonment on the payment step.

Read each session note, tag it as a real friction signal or normal behavior, and write the specific fix each real signal points to.

Before you start

What you'll need

Free path (everything below is enough to finish)

Hotjar

FreemiumWatch and tag session recordings for rage clicks, hesitation, and scroll-back patterns

Free tier includes session recordings and heatmaps, sufficient for a single funnel audit

Google Sheets

FreeLog each session's tag (SIGNAL/NOISE), matched pattern, and recommended fix

Keeps the friction inventory in one shareable place

The process

1 step

Step 01 of 01

Watch Session Recordings

The lesson's Step 3 lists specific signals to watch for in recordings: rage clicks, repeated scroll-backs, hesitation before a form field, and cursor hovering over the exit button.

Of these 5 session notes, which ones are real friction signals and which are just normal, unremarkable browsing?

Hotjar— Session recordings library, filtered to the payment step, sorted by duration.

Procedure

Read all 5 session notes end to end before tagging any
Tag each as SIGNAL (matches a Step 3 friction pattern) or NOISE (normal behavior)
For each SIGNAL, name the specific pattern it matches and the fix it points to

Sample output

Session 1: User clicks the 'Apply Coupon' button 7 times in 4 seconds after it visibly greys out. Duration 38s.
Session 2: User scrolls the payment page top-to-bottom twice, pauses 12s on the delivery-fee line, then closes tab.
Session 3: User reads menu, adds 2 items, checks out normally in 90s, no unusual behavior.
Session 4: User hovers over the browser back button for 6s while the address field is empty, then fills it and continues.
Session 5: User scrolls smoothly through the order summary once, taps 'Place Order', done in 45s.

Healthy

Sessions 1, 2, and 4 get tagged SIGNAL: rage click on a dead button, hesitation over an unexplained fee, and cursor-toward-exit while stuck on a required field. Sessions 3 and 5 get tagged NOISE.

Unhealthy

Tagging all 5 sessions as friction because the reviewer assumes every recording in the 'high drop-off' segment must show a problem.

What this means

Most sessions in any drop-off segment are unremarkable. The audit's value comes from correctly separating the few real signals from the majority of normal browsing, not from finding a problem in every recording.

So what do I do about it?

Symptom	Action	Effort
The 'Apply Coupon' button greys out with no explanation while still appearing clickable	Add a disabled visual state and inline message explaining why the button is inactive	5 min
Users hesitate on the delivery-fee line before abandoning	Surface the delivery fee earlier in the flow, before the payment step	half day

YouYou can do this yourself, no engineering access required.

Final deliverable

A tagged inventory of all 5 sessions (SIGNAL or NOISE), with the matched friction pattern and a specific fix for each real signal.

See a reference example

Sample output

YETI, payment-step session review (excerpt)

SIGNAL: Session 3, rage click on 'Continue' button during a 2s page freeze. Fix: investigate page load performance on that step.
NOISE: Session 7, normal 60s checkout with no unusual scroll or click patterns.

Success criteria

You're done when you can:

Correctly separates real friction signals from normal browsing across the 5 sessions
Each SIGNAL is tied to a specific fix, not a vague 'improve UX' recommendation