The First Ninety Days of a Programme

Retention measures lag. The first quarter of a programme is mostly about doing the work consistently while the figures have not yet moved.

What moves quickly

Day-two attendance, within weeks.

Week-one completion.

Offer-to-start drop-out, if the contact sequence was introduced.

And the count of logistical failures — starters with no equipment, no supervisor expecting them — which should fall to near zero immediately because they are pure process.

What moves slowly

Ninety-day survival, which by definition takes ninety days to show anything and two quarters to show a trend.

Annual attrition, which takes a year and is contaminated by everything else.

And referral volume, which follows the change in how the place is seen and lags it by months.

Which means the first quarter is judged on the leading indicators or it is judged on nothing.

The weekly rhythm

Starters this week, and whether each had a named person expecting them.

Day-three conversations held against due.

Equipment ready on day one.

Day-two attendance.

Four lines, five minutes, and they are the operational check that the arrangement is actually happening.

The monthly rhythm

Ninety-day survival for the cohort that is now mature.

Tenure curve redrawn.

Exit conversations from the month, grouped.

One change made and communicated.

And the agency figures, discussed with the supplier rather than filed.

The pressure to add things

By week six somebody will propose a recognition scheme, an app, a survey or a bonus.

Which is the point at which programmes lose their shape.

Hold the line: the interventions already chosen, done consistently, until there is evidence about them.

Adding a second thing before the first has been measured means neither can be attributed, and the whole exercise becomes a matter of opinion.

What usually goes wrong in the quarter

The day-three conversation stops happening in week five, because supervisors are busy and nobody is counting.

Which is why it is on the weekly list.

The buddy target adjustment is not applied in the system, so the scheme quietly reverts.

The advert reverts to the old copy when the agency runs a campaign.

And the owner is pulled onto something urgent.

All four are drift rather than failure and all four are caught by counting.

Telling people

At the start: what is changing and why, in plain terms.

At the end of the quarter: what the numbers did, including where they did not move.

Which is the part that determines whether the workforce engages with the next round, and an operation that reports only successes is read accurately.

The review at week twelve

Leading indicators, with before and after.

The first mature cohort's ninety-day survival.

What was done consistently and what drifted.

One honest statement about what has not worked.

And the decision about what to add next, which is now an evidenced decision rather than an idea.

The measure

Interventions delivered against interventions planned, weekly.

Leading indicators, trended.

Ninety-day survival for each cohort as it matures.

And whether the owner still has the time allocated, which is the first thing removed and the thing everything else depends on.

Protecting the quarter

No new initiatives for ninety days.

Which is the hardest part and the one that determines whether anything can be attributed.

Write it down at the start so the request is answered by reference.

How it fails in practice

Four initiatives launch in one quarter, the figures move for unrelated reasons, and nothing can be attributed.

The whole approach is then judged on an opinion.

The check

Count interventions delivered against planned, weekly.

Drift appears in the process measures a quarter before it appears in attrition.