surface
Being interviewed?

Uses

  1. IPeople and culture
  2. IICustomer experience
  3. IIIOperations

For

More

Book a conversation

People and culture

You ran the survey and presented the results, and nothing changed.

Surface talks to your people one to one, by voice or text, while it reads what your systems hold about the same weeks: starters and leavers, absence, rotas. You see where people’s accounts and the data agree, where they part, and what to change first.

Fig. 06, The frieze: A long carved procession along the upper wall of a temple above a calm sea, the figures softened by weather.

What you’re up against

The results arrive. The explanation stays in the building.

  1. 01

    The score with no explanation

    Engagement fell at two sites. The comments are too vague to act on, so the team guesses, then defends the guess.

  2. 02

    The second survey

    People notice when nothing came of the last one, and the next survey gets fewer and shorter answers.

  3. 03

    The same question, every review

    Onboarding, exits, absence, the months after a restructure. Each has its own report, and none of them says what to change on Monday.

Beneath the process

A new starter’s first month

Ask how a new starter’s first month goes and you hear the version drawn here. Move the lens along it to see the same weeks as the systems recorded them.

01Laptop and system access

Reported
Ready on day 1
Observed
Day 9 on average
Prescribed
Day 1

02Buddy assigned

Reported
Before day 1
Observed
After day 14 for a third of starters
Prescribed
Before day 1

03Day 30 review

Reported
Every starter
Observed
Logged for 61%
Prescribed
Every starter

Illustrative. Starters, dates and figures are invented.

One investigation

Two sites, two days’ notice.

Follow one investigation from a sentence in an interview to a decision. The people, sites and figures are invented. The method is how Surface works.

  1. 01 What people said

    “We find out our shifts on WhatsApp the night before. You can’t plan anything.”
    Bar staff, site 4Invented excerpt
  2. 02 What the systems recorded

    Rota publishing times, last six months

    SitesMedian noticeLate changes
    Site 42 days3 times the rest
    Site 72 days3 times the rest
    The other seven12 to 16 daysBaseline

    2 days of notice at the two sites whose scores fell

    The two sites with the lowest scores are the two whose rotas arrive last. Leavers there ran at twice the rate of the other seven.

    Six months of rota publishing times by site, from an export of the scheduling system. Notice is the time between a rota being published and its first shift.

    A gap in notice shows where to look. It does not establish the cause on its own.

  3. 03 A second account

    “Head office moves our labour target every week, and I rebuild the rota every time.”
    Site manager, site 7Invented excerpt
    “Head office moves our labour target every week, and I rebuild the rota every time.”

    A second account points at the labour target. Check the target history before treating it as the explanation.

  4. 04 The procedure

    Rota policy

    The policy says

    • Rotas published 14 days ahead

    At sites 4 and 7

    • Labour target revised weekly
    • Rota rebuilt after each revision

    The policy assumes a target that holds for two weeks

    The policy asks every site to publish rotas fourteen days ahead. It does not say what a site should do when its labour target changes inside that window.

    Open question: who sets the targets for sites 4 and 7, and on what schedule?

  5. 05 The finding

    Finding for review

    Two sites rebuild their rotas every week, because their labour targets move every week.

    Sites 4 and 7 sit next to a stadium and are the only two whose targets are revised weekly, against the fixture list. Their rotas follow the target, so they arrive two days ahead.

    Ruled outThe managers. Both sites changed manager this year, and so did three sites whose scores held.

  6. 06 The decision

    Set the targets three weeks out.

    Set those two sites’ labour targets from the published fixtures three weeks ahead, and publish their rotas ten days ahead.

    Owner
    Operations director, with HR
    Next step
    Targets from the fixture list
    Compare
    Notice and leavers next quarter

An illustrative investigation. People, sites and figures are invented.

Who we talk to, what we read

Everyone who does the work, and what the systems kept.

Surface talks to

  • Staff at every level
  • Their managers
  • HR
  • The people who set targets and rotas

Surface reads

  • Survey results by team
  • Starters and leavers
  • Absence
  • Rotas and shift changes
  • The policies themselves

In the pilot, Surface reads exports of the systems you already keep. What it can show depends on what they hold and how far back. We check that in the first call.

What you come away with

A brief you can act on.

Each finding sets what people said beside what the systems recorded and what the procedure says, names the explanation it ruled out, and points to a decision with an owner.

  • Every finding links to the verbatim quotes behind it.
  • Numbers are calculated from stated inputs, not written by the AI.
  • Findings are checked across three independent runs. A headline that does not hold up is withheld.
Finding for reviewIllustrative

Engagement fell at the two sites whose rotas arrive two days ahead.

Reported
“We find out our shifts on WhatsApp the night before.”
Observed
Median rota notice of 2 days at sites 4 and 7; 12 to 16 days at the other seven.
Prescribed
Rotas published 14 days ahead.

Ruled outThe managers. Both sites changed manager this year, and so did three sites whose scores held.

Decision
Targets set three weeks out
Owner
Operations director, with HR
Compare
Notice and leavers next quarter
Read the full example

Before we begin

Questions people ask.

The page for people being interviewed sets out what they are told.

How are people quoted?

Each person chooses: by name, by role only, or not quoted at all. If they don’t choose, they are quoted by role only.

Who can read what people say?

Your company sees findings and the quotes people allowed, never a transcript. A small number of Surface staff can open a transcript to support the service.

Does this replace our engagement survey?

No. Keep the survey. Surface follows it up with the people and the data behind the scores.

What do you need from our systems?

Exports of what you already keep: survey results by team, starters and leavers, absence and rotas. In the pilot, we work through the exports with you.

The offer

One question, ten people, one brief.

We agree one question about one process with you, interview up to ten of your people by voice or text, read the records that go with it, and take you through a brief that shows the evidence for every finding.

Book a conversation

Consultant, assessment firm or CX agency? Work with Surface

Illustrative evidence

Evidence

Invented material, written to show how evidence connects.