Skip to content

Surface-Sampling Walkthrough

Field walkthrough method — instantiates Latent Affordance Surfacing

Walks the intended user's real journey in the first person, surface by surface, recording where attention actually lands and where people are pulled — the map of where a cue would be seen and where it would be missed.

You cannot place a cue where users look until you know where they look. Surface-Sampling Walkthrough is the field method that finds out: an evaluator role-plays the intended user and moves through the actual ordered sequence of surfaces they encounter — screens, pages, rooms, emails, moments — narrating at each stop what is noticeable and where the natural pull to go next leads. Its distinguishing move is that it is embodied and sequential rather than a desk cross-walk: it does not hold two lists and mark cells, it travels the path and watches attention behave, producing the ordered sampling map and a record of the desire paths — the routes people are drawn down whether or not a designer intended them. That map is the substrate every downstream surfacing decision stands on.

Example

A public library wants patrons to use services that already exist — study-room booking, free museum passes, an ask-a-librarian chat — that almost no one touches. Before designing any signage, a librarian runs a Surface-Sampling Walkthrough as a new patron: front-door signage, then the website homepage, then the catalog search, then (only if pushed) the account page. At each surface she records what a first-timer would notice and where she feels pulled to go. The finding: the homepage — sampled by nearly everyone — pulls patrons straight into catalog search and says nothing about rooms or passes, while room booking sits on an account page perhaps one patron in twenty-five ever opens.

The deliverable is an ordered map of sampled surfaces annotated with the desire paths at each — "from the homepage, everyone reaches for the search box." It tells the surfacing work not just that a capability is buried, but which high-traffic surface and which natural pull a cue should ride, and where a cue would simply never be seen.

How it works

  • Adopt the user's role and start where they start. Pick a real persona and enter at their true first surface, not the sitemap's front door; move only where that user would actually move.
  • Narrate attention and pull at each surface. For every stop, record what is perceivable and — the distinctive part — where the natural next-action pull leads, capturing the desire path even when it runs away from the buried capability.
  • Assemble the ordered map. Sequence the sampled surfaces and weight them by how commonly that path is taken, so later work can target the surfaces and moments users genuinely reach.

The walkthrough is first-person and in-situ: its value is catching the lived pull of a journey — the thing a static inventory of screens cannot feel.

Tuning parameters

  • Persona and entry point — whose journey, starting where. The wrong persona or an idealized entry point maps a path no real user walks.
  • Surface granularity — every screen and doorway, or only the decision moments. Fine granularity catches a buried step; coarse granularity keeps the map legible.
  • Grounding evidence — walked from real analytics and observation, or from the team's imagination. Imagined paths map fiction; the honest default anchors each step in evidence of where users actually go.
  • Desire-path sensitivity — how aggressively you record every emergent pull versus only strong ones. More sensitivity surfaces sleeper routes but lengthens the map.
  • Physical + digital scope — whether the walk crosses channels (a sign, then a site, then an email) or stays in one. Cross-channel walks catch handoff gaps but cost more to run.

When it helps, and when it misleads

Its strength is that it stops teams from planting cues on surfaces nobody visits and reveals the high-traffic surfaces and natural pulls that are starved of any signpost — the where that makes later cue design land. It is the cheapest way to replace assumptions about the journey with a walked one.

Its central failure mode is walking the designer's imagined path instead of the user's, which produces a confident map of a journey no one takes; the classic misuse is walking it to confirm that the cue you already placed sits "on the path." It also captures a single evaluator's read of attention, which is a proxy for real behavior, not a measurement of it. The discipline that guards against this is to ground every step in real traffic or observation, include the abandonment points, and treat the map as an input to test — not proof that surfacing will work.[1]

How it implements the components

Surface-Sampling Walkthrough realizes the where-and-toward-what side of the archetype — the terrain map cue design needs, not the cue itself:

  • user_sampling_path_map — its core output: the ordered, weighted map of the surfaces and moments the intended user actually inspects, built by traversal rather than assumed.
  • desire_path_feedback_capture — its distinctive capture: the natural routes and pulls users follow at each surface, recorded in situ so cues can be placed with the current of attention rather than against it.

It does not enumerate the full capability inventory or rank the visibility gap by inspection — that comprehensive cross-walk is Affordance Visibility Audit, which consumes this map; nor does it measure a discovery rate (First-Attempt Discovery Test) or design any cue (Signage and Wayfinding Revision).

Notes

The walkthrough and the Affordance Visibility Audit are easy to confuse and are cleanest kept apart: the walkthrough is the field method that builds the sampling map and feels the desire paths by traveling the journey; the audit is the desk cross-walk that holds that map against the full capability list to rank the gaps. Run the walkthrough first — it produces the where; the audit turns the where into a prioritized worklist.

References

[1] The first-person, step-through form here is kin to the HCI cognitive walkthrough, in which an evaluator plays the user and asks at each step whether the right action is perceivable and obvious. The technique is real and long-established; it is used here only as a means to map sampled surfaces and desire paths, and the library figures are illustrative. ↩