Concepts

Switch: Changing When Change Is Difficult

You’ve explained the new working method three times. Two teams are using it. The rest aren’t. Which question will you put on the agenda on Monday morning: why those four teams aren’t using it, or what exactly those two teams are doing differently? That’s not a rhetorical choice. It determines whether you’ll be analysing or copying over the coming month.

You have explained the new discount rule three times: anything above 15 per cent needs the CFO's sign-off. Two sales pods follow it. The rest do not. Which question goes into Monday's pipeline review: why those four pods work around it, or what those two pods actually do differently? The answer decides whether you spend the next month analysing or copying.

You have explained the new route-planning system three times. Two branches now plan with it; the other four still work from their own spreadsheets. Which question goes on Monday's 7.30 agenda: why those four depots lag behind, or what those two planning desks actually do differently? The answer decides whether you spend next month analysing or copying.

You have explained three times that every change request gets a price before a developer touches it. Two project teams do it. The other four just keep building. Which question goes on Monday's agenda: why those four don't, or what those two actually do differently? That choice decides whether you spend next month analysing or copying.

You have explained three times that every engineering change goes into the ERP as an ECR before assembly sees it. Two project teams do it. The other four do not. Which question goes on Monday's agenda: why those four skip it, or what those two actually do differently? That decides whether you spend the next month analysing or copying.

I’d like my bright-spot ranking and the script

Two hours · senior consultant · bright-spot ranking and script to take back with me

Most change programmes get bogged down because of an incorrect diagnosis

The fundamental attribution error

You observe behaviour that isn’t keeping pace, and you interpret this as a reflection of character: a lack of urgency, resistance, outdated thinking. Chip and Dan Heath refer to this in *Switch* (2010) as the fundamental attribution error. What appears to be unwillingness is often simply a situational problem.

Your CSMs are not running the expansion playbook, and you read character into it: no commercial drive, too cosy with the customer. Chip and Dan Heath call this the fundamental attribution error in Switch (2010). A CSM carrying eighty accounts and a full ticket queue has a situation problem.

Drivers skip logging waiting time at the dock in the app, and you read unwillingness: they don't want to be watched. In Switch (2010), Chip and Dan Heath call this the fundamental attribution error. Often the situation is the cause: gloves on, no signal at dock 14, a shipper shouting.

Your senior developer writes no handover notes and you read indifference into it. Chip and Dan Heath call this the fundamental attribution error in Switch (2010). Look at his week: every incident at your biggest account lands on him. What looks like unwillingness is a problem of situation.

Your service engineer files his report a week after the breakdown, and you read that as sloppiness. In Switch (2010), Chip and Dan Heath call this the fundamental attribution error. Often he is standing in a German factory hall with no signal and a forty-field form on a tablet.

Self-control is a finite resource

The second mechanism comes from the same book: self-control is a finite resource. The replication studies carried out since 2015 have shown this effect to be smaller than *Switch* assumes, so we are using it as a working hypothesis, not as a law.

Any action that requires conscious oversight draws on the same resource. By its very nature, a major change programme requires a great deal of conscious oversight. What you perceive as laziness in week six is therefore simply exhaustion.

Every action that needs conscious oversight draws on the same reserve. An AE who fills in seven MEDDICC fields in Salesforce after every discovery call pays that bill daily. What looks like sloppy CRM hygiene in week six is exhaustion.

Every action that needs conscious oversight drains the same source. A planner who checks the new German toll surcharge by hand on every order does so on top of sixty routine decisions a day. What looks like sloppiness in week six is exhaustion.

Each time a client asks for something small in Slack, your project lead has to choose: absorb it or price it separately. That call takes conscious effort, twenty times a week. What looks like slackness in week six is exhaustion.

Every action that needs conscious oversight draws on the same reserve. A planner who checks each machine's BOM against the ERP every Friday does it on top of his job. What looks like carelessness in week six is simply exhaustion.

Work it out using your own figures.

33%

Adoption rate

First, work out your adoption rate: the number of teams that are actually using the new working method, divided by the total number of teams that should be using it. The denominator is therefore the total number of teams, not the number of teams you have trained. Example: 2 out of 6 teams are using it, so 33 per cent.

Start with your adoption rate: the number of AEs who genuinely use the new forecasting method, divided by every AE carrying a quota. The denominator is everyone with a quota, including the two who started last month, not only those who sat the training. Example: 3 of 11 AEs use it, so 27 per cent.

Start with your adoption rate: the number of branches that actually use the new margin-per-lane report in their weekly review, divided by all branches that should use it. So the denominator is every branch, including those where nobody attended the training. Example: 3 of 8 branches use it, so 38 per cent.

Start with your adoption rate: the number of teams actually running the new way of working, divided by every team that should run it. So count your support and managed services teams too, not only the teams that sat the training. Example: 3 of 9 teams run it, so 33 per cent.

First work out your adoption rate: the number of project managers who actually send the milestone invoice within five working days of the FAT, divided by all project managers with a live project. Not by the three who attended the training. Example: 3 out of 9 do it, so 33 per cent.

180 hours

The sum of the standstill

Then there’s the cost of downtime. Here are some example figures – fill in your own: three announced changes this year, each requiring 60 hours of preparation, at an internal hourly rate of 85 euros. That’s 180 hours and 15,300 euros in preparation costs. If one of the three goes ahead, 10,200 euros will be tied up.

Then put a price on the standstill. Sample figures, use your own: three changes this year (new pricing tiers, annual prepayment, a sharper ICP), each 80 hours of preparation at an internal rate of 95 euros. That is 22,800 euros. If one of the three runs, 15,200 euros sits idle, paid out of your runway.

Then price the standstill. Example figures, fill in your own: three announced changes this year, such as digital consignment notes, dock slot booking and a new fuel clause. Each took 70 hours of preparation at 80 euros internal rate: 16,800 euros in total. If one runs, 11,200 euros stands idle.

Then price the standstill. Example figures, use your own: three announced changes this year, such as a standard SOW, value pricing and a retainer model. Each took 50 senior hours you would otherwise bill at 110 euros. Together 16,500 euros. If one runs, 11,000 euros sits idle.

Then put a price on the standstill. Example figures, use your own: this year a new quotation template, a standard service contract and the BOM in the PLM, each 70 hours of preparation at 90 euros internal rate. That is 210 hours and 18,900 euros. If one runs, 12,600 euros sits idle.

An industrial machine standing still in a workshop, with no one at the controls

Our estimate; not a measured standard.

Every announcement that comes to nothing reduces the uptake rate of the next one, because anyone who has seen it fizzle out twice will wait to see if it’s for real this time. That is our observation amongst clients; it is not a measured figure, and that item does not feature in any post-project analysis.

Every announcement that came to nothing lowers adoption of the next one. Anyone who watched the product-led motion and the previous ICP quietly die will wait for the next board meeting this time. That is what we observe with clients, not a measured figure, and that line appears in no board deck.

A driver who saw the last trip-logging app vanish after two months, like the customer portal nobody opened any more, waits at the next announcement until it proves serious. So every change that fizzled drags down adoption of the next. That is what we observe at clients, not a measured figure, and no trip costing shows that line.

Every announcement that came to nothing weighs on the next. The developer who watched both the promised knowledge sharing and the new estimation method fade out now waits to see if you mean it. That is what we observe with clients, not a measured figure, and it appears in no project post-calculation.

A production planner who watched the digital planning board and the previous ECR procedure fade away will wait at the third announcement to see whether management means it. That is what we observe with clients, not a measured figure. And that cost appears in no post-calculation.

[ Four questions ]

Four questions

Answer them using your own figures or your own observations, not based on a hunch.

adoptiegraad = teams die het draaien / teams die het zouden moeten draaien

  1. What is your adoption rate for the most recent change you announced, expressed as the number of teams running it divided by the number of teams that should be running it?

  2. Which team is currently achieving the outcome you want from everyone, and do you measure that outcome using a tool or based on your impression?

  3. Which step in the new process requires more clicks, more screens or more waiting time than the old one, measured over a single complete cycle?

    Which step in the new renewal flow in HubSpot costs more clicks, screens or waiting for approval than the old one, measured over one full cycle from churn signal to signed renewal?

    Which step in the new TMS order entry costs your planner more clicks, more screens or more waiting than the old one, measured over one trip from order to invoice?

    Which step in the new time booking per task code takes more clicks, screens or waiting than the old weekly timesheet, measured over one full sprint?

    Which step in the new spare-parts quotation takes more clicks, screens or waiting time than the old one, measured from enquiry to dispatch?

  4. At what point in the week does a person receive feedback that they are doing well, and how many days elapse between the action and that feedback?

Switch divides each change into three parts that you can tighten separately

Chip and Dan Heath call them ‘Direct the Rider’, ‘Motivate the Elephant’ and ‘Shape the Path’. Under each section, there are specific components listed, not just general questions.

Direct the Rider

Send the Rider

De Ruiter is the analytical part that seeks direction and keeps mulling things over when the direction is unclear.

Under ‘Direct the Rider’: Find the Bright Spots, Script the Critical Moves, Point to the Destination.

Motivate the Elephant

Move the Elephant

The Elephant is the part that provides the energy and prefers the comforts of the present day.

Under the heading ‘Motivate the Elephant: Find the Feeling, Shrink the Change, Grow Your People’.

Shape the Path

Design the Path

The Path is the environment: the system, the procedure, the physical layout.

Under ‘Shape the Path: Tweak the Environment, Build Habits and Action Triggers, Rally the Herd’.

A portrait made up of individual mosaic tiles, each of which has a different colour when viewed up close

Week one: Find the Bright Spots

What do you do in week one as the person with ultimate responsibility? Just ‘Find the Bright Spots’. Not the failure analysis – that takes a month and identifies causes that you can’t replicate.

The book *Switch* (2010) cites the example of Vietnam: families who fed their children prawns and sweet potato leaves four times a day had healthy children, even whilst living in the same conditions of poverty. The solution was already available locally.

Select along two axes

Rate each team on a scale of 1 to 5 for ‘measurable difference’, where 5 means that the difference is visible in the tool, not just in the impression. Then rate them on a scale of 1 to 5 for ‘replicability’, where 5 means that the action can be carried out without additional authorisation, an additional system or an additional budget. Higher is better on both axes. A team scoring below 4 for measurable difference is eliminated, and a team scoring below 3 for replicability is eliminated. Of those remaining, the team with the highest sum of both scores wins; in the event of a tie, the team with the highest replicability wins. If nothing remains, your first task is not to rank but to measure. This is our rule of thumb, not a standard taken from a book.

From ‘bright spot’ to ‘script’

Don’t ask the chosen team why things are going well, but what they’re actually doing, in terms of actions and timing. That will be your ‘Script the Critical Moves’: a maximum of five lines, each describing an action with a trigger.

Ask the AE with the shortest CAC payback not why she wins but what she does and when, such as sending the mutual action plan within an hour of the demo. That becomes your Script the Critical Moves: five lines at most, each an action with a trigger.

Don't ask the depot that makes it work why it succeeds; ask what its planners do, and when. Say, at 15.30, at order cut-off, they check tomorrow's driving hours against the plan. That becomes your script: five lines at most, each an action with a trigger.

Don't ask the team whose retainer client renews every year why it works. Ask what they do: who calls the client, on which day, after which ticket. That becomes your Script the Critical Moves: five lines at most, each an action with a trigger.

Ask the service team that closes maintenance contracts what they do and when. For example: at commissioning the contract lies next to the handover protocol. That becomes your Script the Critical Moves: five lines at most, each with a trigger.

Week two involves ‘Point to the Destination’ as a single black-and-white objective, plus testing the script with one other team.

A bunch of keys on a ring, one of which is being held separately
[ By sector ]

The same three steps, three different diagnoses

The unit of measurement, the instrument and the clock vary from sector to sector. One rule applies across all three: the ‘bright spot’ belongs to the team that produces it. They decide what goes into the description and they are the first to read it.

SaaS scale-up

  1. 01 Direct the Rider. Unit of measurement: lead time from signed contract to first productive user, in days. Metric: status transitions in your CRM and your ticketing system. Timer: two-week sprint. Identify the onboarding team with the shortest median lead time across at least ten customers. Work with that team to document what happens on day one; they decide what goes into the description and they are the first to read it.
  2. 02 Motivate the Elephant. Shrink the Change: one sprint, one segment, not the entire customer base. During the sprint review, ask the team to show their own lead time curve for the last ten customers, not that of a benchmark. Trigger: the curve is displayed on the screen before anyone starts talking about capacity.
  3. 03 Shape the Path. Tweak the Environment: remove the fields that nobody fills in, and make the first critical action a mandatory status transition in the flow. Anything that isn’t in the flow has to be remembered by someone, and that is precisely the resource that gets used up.
Find out how we approach this metric at SaaS scale-ups

Logistics

  1. 01 Direct the Rider. Measurement unit: number of scan discrepancies per journey and waiting time during cross-docking, in minutes. Tool: scan logs from your TMS and the trip reports. Frequency: daily, because today’s error will be impossible to trace tomorrow. Identify the branch or team with the lowest number of discrepancies per trip.
  2. 02 Motivate the Elephant. Find the Feeling without the drama: let the planners join in on the ride with the most deviations just this once.
  3. 03 Shape the Path. Build Habits and Action Triggers: link the scanning action to a physical moment that’s already part of your routine, such as closing the door or signing out at the reception desk. Rally the Herd via the daily board in the workshop, using team scores rather than individual names.
Find out what scan discrepancies per journey mean in the logistics sector

Mechanical Engineering

  1. 01 Direct the Rider. Unit of measurement: the difference between the preliminary calculation and the final calculation in hours per service order, plus the number of bill of materials changes after approval. Tool: time recording and service reports in your ERP system. Timing: per project phase, as a lead time of several months does not provide a weekly rhythm. Identify the project team with the smallest post-calculation variance across three completed orders. Three orders are too few to rule out noise, so verify the result with a second project manager before replicating anything. The description remains the property of that project team.
  2. 02 Motivate the Elephant. Grow Your People as a means of building competence, not as a narrative of identity: the mechanic who fills in the service report in full is the mechanic whose additional work is actually invoiced.
  3. 03 Shape the Path. Tweak the Environment: condense the service report down to the fields that actually appear on the invoice. In our work, we have found that any field that serves no purpose reduces the completion rate for all other fields.
See how the post-calculation variance affects the mechanical engineering sector

You’re observing work, not people, and the aim is to reduce the number of clicks for the rest.

Please bring

A two-stage decision-making process

The model must not carry out step 4 until you have replied. That is precisely the pattern we refer to in agentic as a ‘gated agent’: a fixed rubric, a strict rejection rule, and a human step in the middle.

Before you use this

Before you paste anything in: leave out customer names, staff names, rates and anything covered by an NDA. This analysis is based on roles and team codes, not on names, and that is not a compromise but, in fact, the very reason why the results can be replicated. For business data, use a business subscription with a data processing agreement; a free consumer account does not provide that basis.

For your works council

One more thing – and this isn’t just a formality. The aim is to ensure that the other teams have fewer clicks and less work to rectify, not to keep tabs on anyone. Put it in those terms to your works council, and make sure you stick to it afterwards.

Rol: je bent veranderanalist. Ik lever teamgegevens op rolniveau, zonder namen.

Ik plak hieronder:
- De werkwijze die overal zou moeten draaien, in een zin.
- De meeteenheid en het instrument waaruit die komt.
- De meetperiode, met begin- en einddatum.
- Per team (A, B, C ...): teamgrootte, rolsamenstelling, score op de meeteenheid,
  en wat dat team feitelijk anders doet in handelingen.

Opdracht:
1. Rangschik de teams op twee assen. Scoor elk team 1 tot 5 op meetbaar verschil,
   waarbij 5 betekent dat het verschil zichtbaar is in het instrument, niet in de
   indruk. Scoor daarna 1 tot 5 op kopieerbaarheid, waarbij 5 betekent dat de
   handeling uitvoerbaar is zonder extra bevoegdheid, extra systeem of extra budget.
   Hoger is beter op beide assen. Een team dat lager dan 4 scoort op meetbaar
   verschil valt af, en een team dat lager dan 3 scoort op kopieerbaarheid valt af.
   Van wat overblijft wint het team met de hoogste som van beide scores; bij gelijke
   som wint de hoogste kopieerbaarheid. Blijft er niets over, dan is je eerste werk
   niet ranken maar meten. Toon beide scores per team in een tabel.
2. Blijft er geen team over, meld dat, benoem welke meting ontbreekt en stop hier.
   Anders: noem het winnende team de bright spot, motiveer met de cijfers,
   in maximaal 150 woorden.
3. Geef mij tien vragen voor dat team. Alleen handelingsvragen met een tijdstip of
   een trigger. Geen vragen naar houding, motivatie of mening.
4. Wacht op mijn antwoorden. Schrijf daarna een Script the Critical Moves van
   maximaal vijf regels. Elke regel is een handeling met een trigger, geen houding.
5. Benoem per regel welk obstakel op het Pad die regel wegneemt.
6. Sluit af met wat je niet weet en welke gegevens ik nog moet aanleveren.

Where this framework does not work

The authors themselves are clear about this in *Switch* (2010). It is deliberately incomplete, designed for memorability rather than comprehensiveness. It breaks down when someone genuinely prefers the current situation; in that case, no path can be designed, as there is a conflict of interest. It assumes little formal power. If you do have that, and a decision is simply a decision, then you have other tools at your disposal. And it is about behaviour, not about a flawed strategy. A poorly executed plan remains a poor plan. And we do not have our own client data that predicts whether your change will succeed, so we do not pretend to.

Read more

I’d prefer to do this as a working session rather than as an introductory meeting.

  • In two hours, we’ll work together to create your bright-spot ranking, with the two axes filled in using your own figures.
  • In two hours we build your bright-spot ranking of your sales pods together, both axes filled with your own CRM data.
  • In two hours we build your bright-spot ranking of branches together, both axes filled in with your own trip and lane figures.
  • In two hours we build your bright-spot ranking together, with adoption and project margin per team from your own PSA.
  • In two hours we build your bright-spot ranking of project teams, with both axes filled in from your own hours and margins.
  • After that, I’ll work with you to write the ‘Script the Critical Moves’ – no more than five lines – for one specific team.

Take those two items with you, even if we don’t do anything else.

I’d like my bright-spot ranking and the script

Two hours · senior consultant · bright-spot ranking and script to take back with me