Why most internal marketplaces plateau at browsing, not moving
THE VERDICT
An internal talent marketplace doesn't change hiring behaviour by existing. It changes behaviour when it changes who gets to say yes. Most deployments stall six months in because they digitise browsing — employees can now see open gigs, stretch assignments, open roles — without touching the actual friction that keeps people stuck: a manager who has no incentive to let a good performer leave. The platform is real and useful. It just can't fix a release problem by itself.
Evidence base: Industry analysis — pattern observed across public case studies, vendor implementation guides, and how these programmes are typically governed. For: HR, Talent Management, People Analytics, and business unit leaders sponsoring internal mobility.
Most internal marketplace pitches sound the same: surface hidden skills, break down silos, let talent flow to opportunity. The demo is compelling because browsing is the easy 80%. Show an employee a feed of open gigs that match their profile, and the tool looks transformative on day one.
The hard 20% is what happens after someone clicks "apply" on an internal gig. Their manager finds out. Their manager was not consulted, was not warned, and is now missing a person for the next quarter with no backfill plan. In almost every organisation, that manager's performance is measured on their team's output, not on how many people they successfully released into the business. So the manager slow-walks the approval, or quietly discourages the move in the next one-on-one, and the platform's usage data shows plenty of browsing and very little movement.
That gap — browsing without moving — is the single most common failure pattern in internal marketplace rollouts. It shows up in the metrics as high login rates and low fill rates, and it's almost never a platform problem.
What the org chart actually requires:
-A release policy that changes what managers are measured on, not just what employees can see.
-A tie-breaker above both managers for mobility disputes, so moves don't die in a stalemate.
-A skills taxonomy that maps to real, gradeable assignments — not resume keywords lifted from a profile field.
-Time-boxed gigs and stretch work visible alongside full role postings, so mobility doesn't require a resignation-level decision to try something.
-A refresh cadence for skills and interest data, because a marketplace full of two-year-old profiles trains people to stop trusting it within a month.
What the tech platform actually provides:
-Matching and search across a much larger pool than any one manager could scan manually.
-A structured, searchable record of skills and interest that doesn't live in one manager's head.
-Visibility into demand signals — which skills are being requested across the business right now.
-Workflow and audit trail for how a move happened, useful once the governance question is answered.
-Speed, once the release policy exists to move fast against.
Read those two lists back to back and the pattern from the skills-architecture piece repeats: the first list is a governance decision the organisation has to make. The second is a capability you can buy. Buying the second in the hope that it will force the first is how you end up with a marketplace nobody trusts.
The gates before you buy
01 — A named release policy, with a business owner. "We support internal mobility" is not a policy; it's a sentiment. A real policy states what a manager's performance conversation looks like when a direct report successfully moves internally, and who enforces it when a manager quietly blocks a move anyway.
02 — A tie-breaker for disputes. Mobility will produce conflict — a receiving manager who wants someone now, a sending manager who wants to keep them. Someone above both needs the authority to decide, and that person needs to be named before the first dispute, not during it.
03 — Assignments the skills data can actually reference. If your taxonomy describes skills that don't map to any gradeable, time-boxed piece of work, the marketplace becomes a wish list, not a matching engine.
04 — Movement below the level of a full role change. If the only visible opportunities are permanent role postings, you've rebuilt your internal job board with better search. Stretch assignments and short rotations are what actually change hiring behaviour, because they lower the cost of trying.
05 — A trigger to refresh the data. Static profiles decay fast. Without a forcing function — a quarterly nudge, a manager check-in, a project-close survey — the marketplace's picture of who can do what goes stale within two quarters.
If you haven't settled the release policy and the tie-breaker, the platform will settle them for you — by quietly becoming a browsing tool that nobody uses to actually move.
The pathway
None of this is anti-technology. Applied to an organisation that has already decided who has to say yes, a marketplace does something no spreadsheet or manual process will: it surfaces opportunity and skill at a scale one manager's memory never could. The partnership works. It just works second, after the governance decision, not instead of it.
So before the vendor demo, before the taxonomy workshop: who currently has to say yes when someone wants to move — and what happens to that person's own numbers when they say it?