By Ritvars Mētra, Founder of ReadinessCompassThe readiness work lands. Forty-odd findings, a set of RAG statuses, a slide headed “Recommendations” containing phrases like strengthen manager engagement, improve data quality and increase communication.
Everyone agrees it is a good piece of work. Six weeks later, almost nothing has moved.
This is not usually a failure of will, and it is rarely a failure of analysis. It is a failure of translation. Evidence describes a situation; a plan commits named people to specific acts by specific dates. Those are different artefacts, and producing the first does not produce the second.
This is how to do the translation properly, and why the 30/60/90 structure works when it is built correctly — and quietly fails when it is not.
Why findings do not become actions on their own
Three things go wrong, and they go wrong together.
Recommendations are written as aspirations. “Strengthen manager engagement” is a direction of travel, not an act. Nobody can do it on Tuesday, and nobody can tell whether it happened.
Ownership is assigned to a function. “HR” or “the PMO” owns the action, which means no individual wakes up accountable for it.
Dates float. The plan is dated from the day the report was delivered rather than from the milestone the work exists to protect, so the deadlines have no consequence attached.
Fix those three and most readiness evidence converts into movement. Leave any one unfixed and the plan degrades into a status report about itself.
The translation chain
Every finding worth acting on passes through the same seven steps. Skipping any of them is where plans lose their grip.
Finding → risk → decision → action → owner → date → evidence of done.
Worked through with a real example:
- Finding: 62% of finance users report low confidence in the new period-end close, and unaided completion in testing was 55%.
- Risk: the first close after go-live overruns, requiring manual intervention and delaying reporting.
- Decision: do we rehearse the close before cutover, or accept the risk and staff a heavier hypercare?
- Action: run a full mock close with production-volume data for the whole finance team.
- Owner: the Financial Controller. Not “Finance.” A person.
- Date: completed by 14 days before cutover, so there is time to act on what it reveals.
- Evidence of done: mock close completed within eight days with unaided completion above 80%.
Notice that the decision step is separate from the action. Most plans jump straight from finding to action, which hides the fact that a choice was made — and choices that were never explicitly made are the ones that get quietly reversed later.
Notice too that “evidence of done” is not “mock close held.” Holding the rehearsal is activity. The 80% threshold is the outcome the rehearsal existed to produce.
Why specificity is the whole game
The gap between “strengthen manager engagement” and the example above is not pedantry. It is the difference that behavioural research has measured repeatedly.
Locke and Latham’s 35-year programme of goal-setting research established across hundreds of laboratory and field studies that specific, difficult goals produce higher performance than vague exhortations such as “do your best.” Vague goals let people grade themselves generously; specific ones define what success is and remove the ambiguity.
Their theory carries a caveat worth transferring directly into your plan: the effect holds only where the person has sufficient capability and genuine commitment. An ambitious action assigned to someone without the authority, skill or bandwidth to deliver it is not a stretch goal. It is a plan with a known defect.
The second finding is more directly useful still. Gollwitzer and Sheeran’s meta-analysis of implementation intentions — 94 independent tests, more than 8,000 participants — found that specifying the when, where and how of an action in advance, in an if-then form, produced a medium-to-large effect on goal attainment (d = .65). Crucially, that gain was measured over and above already having the goal. The participants in the control conditions wanted the same outcomes. They simply had not specified the trigger.
Applied to readiness actions, this is a small change with a large return. Instead of “managers to reinforce the new process,” write the trigger in:
“In the Monday team meeting, each manager reviews the exception queue from the new system and names one blocker to escalate.”
Same intent. One is a hope; the other has a time, a place and a first move — and it is the reason manager reinforcement either happens or does not.
Why 30/60/90 — and the trap in it
The 30/60/90 structure is a practitioner convention rather than a research finding, and it is worth being honest about that. What makes it useful is not the arithmetic. It is that the three windows hold genuinely different kinds of work, which stops a plan from being an undifferentiated list sorted by enthusiasm.
The trap is counting from the wrong point. Most plans start the clock on the day the report is presented. That date is arbitrary. Anchor the horizons backwards from the milestone the work exists to protect — cutover, wave two, year-end — and the windows acquire meaning, because now a slipped 30-day action visibly eats into the 60-day one and the consequence is legible to a steering committee.
If go-live is seven weeks away, you do not have a 90-day plan. You have a 20/35/49-day plan, and pretending otherwise is how organisations discover in week eight that the important work was scheduled after the event it was meant to influence.
What belongs in each window
| Window | Purpose | What belongs | Evidence of done |
|---|---|---|---|
| 0–30 | Decide and start the slow things | Governance decisions; anything with a long lead time; stopping active harm; assigning owners; the items that block others | Decisions minuted; long-lead work started; owners named and accepted |
| 31–60 | Build and fix | Process redesign, data remediation, rehearsal, targeted practice, closing capability gaps | The fix exists and has been tested with real users |
| 61–90 | Verify and embed | Re-measurement, ownership transfer to the business, embedding in BAU routines and reporting | Metric moved; named business owner in place |
The first window is the one most often misused. It fills up with visible, satisfying activity — workshops, comms, training sessions — while the items with eight-week lead times sit in the 60-day column where they cannot possibly finish in time.
How to decide where an action goes
Not by importance. Sorting by importance puts everything in the first thirty days, which is the most common way these plans collapse in week two.
Sort by four mechanical tests instead:
- Lead time. How long does this take once started? Long-lead items go early regardless of how important they feel, because starting date, not importance, determines whether they land.
- Decision deadline. When is the last responsible moment to choose? Anything gating other work is a 30-day item by definition.
- Dependency. What must exist first? You cannot rehearse a process that has not been redesigned.
- Cost of delay. What does waiting a month actually cost? Often nothing — and admitting that frees capacity for the items where it costs a great deal.
Run those four tests and the sequencing largely decides itself, which also makes it defensible when a sponsor asks why their item is in the second window.
One named human, every time
The single highest-yield discipline in the whole plan is refusing to let a function own an action.
The underlying mechanism is well established in social psychology. Darley and Latané’s classic experiments on diffusion of responsibility found that the more people who could plausibly act, the less likely any individual was to do so, and the longer they took when they did. That research studied bystanders in emergencies rather than transformation programmes, but the mechanism is general: shared responsibility reliably reduces individual responsibility.
“The PMO will monitor adoption” is a sentence engineered to produce exactly that outcome. Three rules follow:
- A person, not a team. One name per action, even where several people do the work.
- Accepted, not assigned. An owner who was not in the room and has not said yes is not an owner. Confirm it explicitly.
- Business, not project, wherever the action outlives the project. PMI’s research on sustaining benefits beyond the project makes the point that post-project work needs the same ownership and attention as delivery work. An action owned by someone who demobilises in six weeks has an expiry date built into it.
This is also the mechanism by which benefits leak after go-live: not because actions were wrong, but because their owners left and nobody inherited them.
Keeping the plan alive
A 30/60/90 plan is a living instrument for roughly a quarter, and it needs three things to stay one.
A standing review, short and specific. Fortnightly, fifteen minutes, three questions per action: is it on track, what is blocking it, does the date still hold? Not a status presentation.
A route for the blocked ones. Most stalled actions are stalled on a decision someone else has to make. If there is no escalation path that resolves within a week, actions will sit at 80% complete indefinitely. Anything that cannot be resolved in the plan’s lifetime belongs in the RAID log rather than quietly rolling forward each fortnight.
Permission to close things. Some actions stop being worth doing. Closing one deliberately, with a reason, is healthy; carrying it amber for three months is how the plan loses credibility and, with it, the attention of the people who could act on the parts that still matter.
Five failure modes worth checking for
Before you circulate the plan, read it against these:
- Everything is in the first thirty days. You sorted by importance rather than by lead time.
- An owner is a function. Nobody is accountable.
- “Done” is an activity. Sessions delivered rather than a threshold reached.
- Dates run from the report, not the milestone. The plan is not connected to the thing it protects.
- Every action is a communication or training action. A real readiness plan changes process, data, decision rights and routines too. If it does not, the diagnosis was probably shallower than the report suggested.
The point of the plan
A readiness assessment that produces no change in behaviour has cost money and bought reassurance. The plan is where the evidence either becomes decisions or becomes a document.
None of the discipline above is sophisticated. Specific actions rather than aspirations. One named human. Dates anchored to the milestone. An outcome that says whether it worked. A short review that people actually attend.
It is unglamorous, and it is the step where most readiness work is lost — not in the analysis, but in the fortnight after it, when everyone agreed the findings were excellent and nobody wrote down who was doing what by when.
For the evidence side of this, see how a diagnostic sprint differs from a readiness survey and how to turn survey findings into leadership action. The change readiness resources cover the wider picture, and the AI-assisted diagnostic toolkit holds the tools these plans usually draw on.