Alternative channels need functional parity, not token retention

What it is

Keeping a staffed counter that handles only a slice of what the digital channel handles — "this transaction is online only", "the counter accepts documents but processes nothing" — is not an alternative channel but a token channel: a non-digital entrance in form, with a shrunken service list in substance. Parity means the same list of transactions completes through either channel, with no surcharge for going in person.

The test is simple: if a person who will only ever visit the counter cannot complete everything the app can complete, there is no alternative.

Why it happens

Parity has three dimensions: functional parity (same coverage of transactions), cost parity (no extra time, fees or documents for the offline path), and access parity (comparable opening hours and travel distance). Token retention usually breaks on the first dimension, then uses surcharges on the second — queues, extra certified copies, repeat visits — to wear down the residual demand as well.

Why does parity not emerge on its own? Because institutions carry a hidden incentive to let the alternative wither: online migration is the headline metric of digitalisation programmes, and the better the offline channel works, the slower the migration. Absent external constraint, a decaying alternative channel is the natural outcome of rational strategy — which is exactly why parity must be a requirement, not a benevolence.

The UK's experience is instructive: the "digital by default" strategy was paired from the start with assisted digital arrangements — support channels for people who cannot get online to complete the same transactions — making the alternative a structural part of the policy rather than an afterthought.

Studying it

• Dual-track evaluations: for the digital and alternative channels separately, tabulate transaction coverage, completion rates, average handling time and number of visits required, then compare. The UK's assisted digital evaluations are a public exemplar of the paradigm.
• Migration impact assessments: around the closure of parts of the offline offer, track refusal rates and the composition of complaints on the alternative channel.
• Paired walkthroughs: send two batches of testers through the same transaction, one per channel, recording completion and cost — turning "parity" into a measurable difference in pass rates.

One methodological caution: measure parity per transaction, not per channel. A matrix of "can transaction X be completed in channel A" is required; channel-level aggregate completion rates hide the missing items.

Where it stops holding

• Parity lives in the citizen's experience, not the institution's cost base. It is true that staffed channels cost more per transaction than online ones; that is not an argument against parity. Parity demands "can it be completed", not "both channels equally cheap".
• A few transactions cannot be fully mirrored physically. Those depending on live queries (real-time balances, instant verification) can only approximate at a counter — such asymmetry must be declared with a compensating measure (callback, next-day response), not silently downgraded.
• Parity is dynamic. Every new transaction launched online adds a row to the parity matrix; align once and it re-decays within months — parity is a standing audit item, not a one-off project.

Applying it

• Publish a side-by-side channel matrix: for every transaction, its availability through the digital and alternative channels, with reasons and workarounds marked where it is missing.
• Before removing any transaction from the alternative channel, run a public assessment: how many are affected, and how those people are served instead.
• Treat alternative-channel usage volume as a monitoring indicator of exclusion scale — a number to read, not to eliminate.
• How to check: sample transactions from the master list at random, send testers through the alternative channel only, and compute the completion gap against the digital channel. Every sampled "cannot complete" is a hole in the parity matrix.

Related

• Same group: Z8.05.1 Services that presuppose a smartphone exclude non-users by design · Z8.05.3 Exclusion concentrates among older, low-income and disabled people · Z8.05.4 Alternative entrances need to be as prominent as the main one, not hidden in a corner
• Nearby: Z8.02 Touchless services and public terminals · Z8.06 Accessibility requirements for public facilities
• Search terms: assisted digital · channel parity · digital by default · service equality