TELECOM · PLANS, BUNDLES & HARDWARE
Reduce plan and bundle abandonment.
Telecom sells a two-year relationship with a device attached. The buyer is comparing an allowance they cannot evaluate, a network they cannot test, and a handset price split across a contract — three separate uncertainties presented as one price.
Plan grid → handset selection → contract term and upfront split → coverage check → checkout or port-in. Loss clusters at the coverage check, which can end the journey outright, and at the point where the monthly figure changes because of the device.
THE FRICTION MIX
Coverage is a hard stop. Everything else is negotiable.
Telecom is the one sector where a single check can legitimately end the session with a no. Treating that no honestly is worth more than any nudge — it stops the buyer wasting ten minutes and stops you paying to acquire a complaint.
Coverage friction
Address, network availability, indoor signal, install date for fixed line. A constraint, not a doubt — and the answer is often genuinely no.
Choice friction
Four plans differing on an allowance nobody can convert into real-world use. Data in gigabytes is a spec; “enough for your streaming” is a decision.
Commitment friction
Twenty-four months plus a device balance. The upgrade path, the early-exit cost, and what happens if they move house.
SIGNALS THIS SECTOR PRODUCES
A repeated coverage check with an edited address is the highest-value signal in this sector: it is unambiguous, it is cheap to resolve, and getting it wrong costs a real installation visit.
The friction vocabulary →A WORKED DECISION
One session, on the plan grid.
01 · WHAT THE SESSION DID
Loops between two plans three times, opens the data-allowance FAQ, toggles the device between 24 and 36 months, and never touches the coverage checker.
02 · THE HYPOTHESIS
Choice friction, confidence 71%. The buyer cannot tell what the allowance difference means in practice; the device toggling is a symptom of trying to make the monthly figures comparable.
03 · THE RESPONSE
Level 1 — passive adaptation. The allowance difference is restated in consequences, in page flow, next to the two plans being compared. No overlay, nothing to dismiss.
WHAT THIS SECTOR USUALLY DOES INSTEAD
“Offer ends midnight — order now”
Add a third recommended plan to the grid
Pop a chat window after 30 seconds
Adding an option to a buyer already stuck between two makes choice friction worse — recommendation surfaces are blocked from firing into a detected choice-friction state. Time on page is not friction.
AND WHERE WE STAY SILENT
A buyer who has checked coverage, chosen a plan and is filling in details gets nothing at all. Nor does one comparing your grid against a competitor tab: that is rational shopping, and no message on your site wins it.
L0 IS STILL THE MOST COMMON DECISION IN THIS SECTOR
THE ECONOMICS HERE
The device subsidy is where conversion-rate optimisation quietly loses money.
Steering hesitant buyers onto the plan with the heaviest handset subsidy raises conversion and destroys contribution. In a sector where margin varies enormously between plan and device combinations, reporting anything other than margin is close to meaningless.
Which is also why we split margin by plan and by device in the pilot definition, and agree it with finance before week one.
WHAT A PILOT SCOPES TO
Two weeks in shadow mode first: every decision recorded, nothing rendered to a single visitor.
How a founding pilot works →THE CASE STUDY WE WANT FROM THIS SECTOR
Reduce plan comparison abandonment without raising the subsidy.
FOUNDING PILOT SLOT
OPEN
The publishable finding here is the one operators argue about internally: how much of plan-grid abandonment is genuinely price, and how much is an allowance nobody has translated into ordinary language. Shadow mode answers that in two weeks without changing a single page.
Bring us your plan grid.
Two weeks in shadow mode, one journey, one outcome. You see the friction distribution for your own traffic before anything renders.