The Kept-Commitment Method
Kinetry's published procedure for commitment attribution: coaching commitments recorded in one cycle, compared against how those exact behaviors moved by the next scored cycle. This page is the full method — including what it cannot claim.
Most coaching and performance tools stop at the commitment: an action item is written down, and whether anything changed is never measured. Kinetry measures the same 24 behaviors every cycle anyway — so it can close the loop: record what people committed to work on, record whether they followed through, and compare how the committed behaviors actually moved against the ones nobody committed to. That comparison is commitment attribution, and the procedure below — the Kept-Commitment Method — is how Kinetry computes and states it.
The method, step by step
During coaching — a 1:1, an advisor session — a person commits to work on one of the 24 defined behaviors, and the commitment is recorded against that exact behavior in the current measurement cycle. Only real commitments count: an AI-suggested action becomes a commitment when the person adopts it, never before, and a note with no specific behavior attached is excluded — what can't be tied to a measured behavior can't be tied to its movement.
When the commitment is reviewed — usually in the next 1:1 — a person marks it kept: the person themself (the checkbox on their own scorecard is the default flow), or whoever coaches or oversees them. Kinetry never infers follow-through from activity signals, calendars, or messages; "kept" means a person said so, on the record. Anything not marked kept by the next scored cycle — dropped with a stated reason, or simply left open — counts as not kept.
Only commitments made in the PRIOR scored cycle count toward the current comparison — the time between the two measurements is theirs to be acted on. A commitment made this cycle is never compared against this cycle's scores, so the ordering is always commit first, measure after — not a same-cycle correlation.
At the next scored cycle, every (person, behavior) pair with a published score in BOTH cycles lands in one of three buckets: committed and kept, committed but not kept, or no commitment made. The comparison is the average score movement of each bucket — kept versus not kept is the follow-through contrast; uncommitted behaviors are the backdrop.
The two-sided contrast renders only with at least 5 (person, behavior) pairs on EACH side; with a thinner not-kept side, only the one-sided line — kept behaviors moved +X — can render, and it needs 5 kept pairs of its own. The person-level contrast — people who kept at least 60% of their commitments versus those who didn't — needs at least 3 people per side. Below those floors the aggregate surfaces stay silent rather than print a number a skeptic would dismiss. Your own private “Since last cycle” ledger is the one exception by design: it is your personal record, not a statistic. Behaviors whose scores were suppressed by the fairness rules in either cycle are excluded entirely.
What the method cannot tell you
These limitations are part of the method, not footnotes to it. Kinetry's wording everywhere — in the product, in reports, on this page — is constrained by them.
The method establishes ordering, not causation. "Behaviors people committed to and kept moved +X" is exactly what Kinetry claims — that the kept commitment came first and the movement followed. It is an observational contrast over real work, not a controlled trial, and every surface that shows the numbers says so.
People who keep commitments differ from people who don't in more ways than the commitment — energy, workload, manager attention. Some of the contrast between the buckets belongs to those differences, not to the coaching. A randomized design could separate them; a measurement product embedded in real teams cannot, and pretending otherwise would be the kind of claim this page exists to rule out.
Each comparison spans exactly one cycle transition, and inside a single workspace the buckets can be small — which is why the floors above exist. The numbers shown are dated with the cycles they compare and are never pooled across companies.
"Kept" is marked by a person — often the person themself — and people are generous with themselves and each other. The method's defense is transparency, not omniscience: commitments are created on the record, carry a named stamp when they're discussed in a 1:1, and a workspace that games its own follow-through only defeats its own measurement.
- • The executive dashboard's “Did the coaching move behavior?” card — the aggregate three-bucket contrast, dated with the two cycles it compares.
- • The Cycle Impact Report's attribution line, in the same words.
- • Each person's “Since last cycle” view — their own kept and dropped commitments against their own behavior movement.
- • Follow-through statistics — the share of recorded commitments kept, counted across all commitments (a broader set than the behavior-tied subset the method compares), shown beside the method's output.
Related definitions: the fairness rules that decide when a behavior score exists at all, the 24 behaviors themselves, and the team behavior instrument category. The Kinetry Pair Model — the published pair taxonomy behind the Team Relationship Map — is the method's sibling artifact, and the product-wide limits are stated in full at What Kinetry cannot tell you.