River
Y CombinatorBacked by Y Combinator
FREE TEMPLATE

Customer Onboarding Plan Template

Four sheets that set the finish line, the clock and the length from the customer's side, then count adoption by named human being.

Free download  ·  No account needed

Value Definition · Kelbrook Distribution · agreed at kickoff

“We know it's working when a depot manager publishes next week's rota on Thursday afternoon without ringing anyone.”

Their words, from the kickoff recording. Every clause below is load-bearing.

ClauseWhat it rules out
A depot managerNot their administrator, and not our implementation consultant
Without ringing anyoneNot a rota built in a training session, and not one that needed a call to head office
Thursday afternoonTheir operating cycle. Rotas go out Thursday so people know their week by Friday

Two events, two dates

Our activation event, first rota createdDay 4, by our consultant, in a sandbox
Their event, witnessed by D. FarrarDay 58, by S. Iqbal, unassisted
Their own deadline, peak season opensDay 62

Four days of slack. Measured from signature this reads as a four-day onboarding, which is a number nobody should say out loud. Measured against the date they published to their own customers, it is a plan that only just fitted.

Illustrative figures for one fictional workforce scheduling vendor and one of its customers.

Page one has this solved and every part of the answer is set by you. Four phases from pre-kickoff to a thirty sixty ninety review, an owner on every milestone, a defined activation event. Time to first value is measured from contract signature and tracked by cohort, so a slow account cannot hide in an average. It is a good build. Three things in it belong to the customer and get held by the vendor anyway.

The finish line is yours, and your activation event usually happens on day four in a sandbox with your own consultant driving. What the customer would call working is a different event, in their words, with a witness. The clock is yours too. Their date is a real thing in their business, and for a federal customer it is set by statute: the fiscal year ends 30 September whatever your kickoff date was.

And the length is asserted. What decides it is eight customer facts, all knowable in week one. How many systems the data sits in, whether a named admin has time allocated, whether their security review has started, whether the incumbent will release the data. In the UK and EU a customer can require that in a structured, machine-readable format. Each factor costs days you can look up in your own history. Adoption then gets counted by person rather than by account, and the quarterly review inherits something that works.

Fifty-nine days, and where every one of them came from

The readiness read that derives the plan length from customer-side facts, the adoption sheet that counts humans rather than accounts, and the history the prediction is calibrated against.

Readiness Read

Eight customer-side facts, asked out loud in the kickoff session rather than sent as a form afterwards. Days come from Merrow's own history, not from judgement.

FactorTheir answerStateDaysBasis
Security reviewNot started. Infosec lead back 22 SeptemberBlocker+18median of 5
Where the data livesThree systems plus two spreadsheets, one of which is the real oneFriction+11median of 4
Users in the roomNo. Bought centrally, used in six depotsFriction+9median of 6
Named administratorD. Farrar, one day a weekFriction+6median of 3
Integrations before valueOne, payroll. The other three are wanted, not neededClear0scoped down in week 1
Their own deadlinePeak season opens 3 November. Published to their customersClear0anchors the plan
Incumbent contractEnds 31 January, notice already servedClear0no dual running
Getting the data outCSV export tested on 200 records in week oneClear0two field mismatches found early

21 base, plus 18, 11, 9 and 6. Fifty-nine days against sixty-two to their deadline. Three days of slack, known in week one rather than in week six, which is what moved the assurance package to day two and scoped three integrations out of the plan.

Every mitigation names the side that can act on it. Four of the eleven critical-path items are the customer's, including the two longest.

Adoption by Person

One row per human being, not one row per account. Account-level activation reads this account as activated.

PersonRoleCore taskUnassistedTrainedState
D. FarrarOps AdministratorYes4YesIndependent
S. IqbalDepot Manager, NorthgateWith help1YesAssisted
P. NakamuraDepot Manager, RiversideNo0YesTrained not started
R. DelgadoDepot Manager, KingsmeadNo0YesTrained not started
L. OkaforDepot Manager, FairfieldNo0NoNot started
B. WhitcombeDepot Manager, EastwayNo0NoNever invited
T. AwuahDepot Manager, LynwoodNo0NoJoined in September
M. ColeOperations Directorn/a0n/aViewer

One of seven operational people can do the task alone, and she is allocated one day a week. This account is one resignation, one illness or one busy Tuesday from zero, and no account-level activation metric can see it.

Never logged in has three different causes here and two of them are the vendor's fault. B. Whitcombe was left off the invitation list and T. Awuah joined after the rollout plan was written. A third group training invitation solves neither.

Onboarding Cohort History

Every completed onboarding with its readiness profile, what was predicted, what happened, and one sentence on what actually cost the time.

CustomerSourcesReview at kickoffPredActualVarTheir dateWhat cost the time
Bexley Freight1Complete2421-3noneNothing. This is the base case
Ashgrove Group2Complete2926-3MetUsers were in the room and it shows
Halloway Rail3Not started5954-5MetReview cleared in 9 days because their deadline was real
Torrance Estates4Not started6166+5noneReview ran 31 days. No deadline meant no pressure to start
Pinevale Federal3In progress5258+6MetContracting officer changed. The statutory date never moved
Northmoor Facilities5Not started7491+17noneThree non-blocking integrations scoped as blocking

The same profile, nine days or thirty-one. Halloway and Torrance arrived identical on paper. The only difference was that Halloway had a real date published to their own customers, so their infosec lead cleared the review in nine days instead of thirty-one.

Kelbrook's comparables are Halloway, Torrance and Northmoor: 54, 66 and 91 days. Northmoor's cause is known and does not apply here, so the working range is 54 to 66 and the sheet says the sample is three.

What's in the pack

01

Value Definition

The customer's own sentence about what working looks like, taken apart into an observable with a named witness, beside your activation event and its real date.

02

Readiness Read

Eight customer-side facts asked out loud at kickoff, each with a state, the days it adds, the basis for those days, a mitigation and the side that can act on it.

03

Onboarding Plan

The length with its arithmetic shown, the comparable onboardings named with their count, the critical path, and what was deliberately moved past the customer's date.

04

Milestone Tracker

Every milestone carrying a side, a named person and days before the customer's deadline rather than days elapsed since a signature nobody outside your company cares about.

05

Adoption by Person

One row per human being, with five states that separate never invited from trained and not started from reluctant, because those go to different people.

06

Onboarding Cohort History

Every past onboarding with its readiness profile, predicted days, actual days, and one sentence on what really cost the time. This is what future plans are derived from.

07

Readiness Standard

How the eight factors are scored, why a blocker is defined by what stops rather than by how annoying it is, and how to test whether a stated deadline is real.

08

Escalation Path

Two named people per side per level, agreed at kickoff when nobody is annoyed, with stated triggers instead of somebody's patience running out.

How to use it

  1. 1

    Open in River, or take it blank

    Open the pack in River and hand it your handoff and kickoff, or download the Word documents and CSV sheets and fill them in yourself.

  2. 2

    Take the readiness read

    Eight questions in the kickoff session. Ask for a name and an allocation rather than a role, and for the date their business actually runs on.

  3. 3

    Derive the length

    Base plus each factor's cost from your own history, with the sample stated, then run the plan backwards from the customer's date.

  4. 4

    Count people, not accounts

    After first value, check how many humans can do the task alone. One is a state most accounts are in and no account metric shows it.

Frequently asked questions

Is this template free?

Yes. Four Word documents and four CSV sheets, downloaded as a zip, no signup. River is the optional half: it takes the readiness read from your kickoff, derives the length from your history, and writes the value definition. More in the template library and the tool index.

We have no history of past onboardings. Does the length still work?

The read still works and the arithmetic waits. Every factor cost is labelled uncalibrated until you have comparable rows, and the plan says so rather than presenting a guess as a forecast. Eight rough rows from memory, with rough durations, is enough to start deriving instead of asserting.

What if the customer has no deadline of their own?

You record the absence, which is itself the strongest predictor in the sheet. In the worked history two identical customers took nine and thirty-one days to clear the same security review, and the only difference was that one had a real date published to its own customers.

How is this different from the handoff from sales?

The handoff is an input here and this pack never writes one. It carries the deal context, the commitments made on calls and the requirement map. This space starts at kickoff and owns the plan, the finish line and adoption, then hands to the quarterly review.

Why not just use our activation event?

Because it is a thing your product does, chosen by you, and it usually fires on day four in a sandbox with your own consultant driving. Both events are recorded here with both dates. The contrast between them is the most useful line in the document.

Isn't per-person adoption tracking overkill for a B2B account?

It is the difference between activated and safe. One administrator who can do the core task on one day a week reads as green at account level and is one resignation from zero. It also matters at renewal, which is where the risk brief picks it up.

The customer says thirty days. What do we do with a derived fifty-nine?

Show them the components. Eighteen days are their security review and eleven are their data being in three places, and both are things they can change. A total is something to argue about, while a breakdown is something to act on, which is why the plan shows its arithmetic.

Derive the plan instead of declaring it

Take the Word documents and CSV sheets blank, or open this exact pack in River and let it read your kickoff and size the plan first.

Edit with AI