Say you’re looking at a replacement for a tool you’ve used for years. You know the shortcuts. You have useful templates. You can finish familiar tasks quickly. Starting again looks like a lot of work.
Some of that work is a real cost still to come. But some of your hesitation may just be about the effort you’ve already put in. Pulling those two apart gives you a better question: which option will serve your work from here?
This module is for working adults reviewing a tool they already know well, who have permission to change it and time to test a realistic alternative.
What you’ll get from this module
In this module
By the end of this module, you’ll have a clearer way to decide whether to keep, fix, test or replace a familiar tool. We’ll look at an economics paper, two decision experiments, an official investigation, a report from regulators and an announcement that a product would be retired. You’ll use what they show to compare the work ahead and pick a situation from daily life. Then you’ll write an agreement that says when to look at your tool again and what must be true before you switch.
Ask Jing
Become a Jingius.
What makes a tool worth keeping?
An economics paper: changing tools can mean real new work.
Joseph Farrell and Paul Klemperer’s 2006 working paper looks at switching costs and network effects in markets. Switching costs are the costs that come with changing suppliers. They arise when a change means paying for something again, such as learning a new product or building a new working relationship.
This matters when you already know a tool well. The hours you spent learning it are gone. But the skill those hours gave you can still help with the work ahead. Replacing that skill may mean new training, or setting up useful arrangements again.
Why it matters: The paper builds economic theory and reviews other research. It isn’t an experiment on how people choose their own software. What it offers here is a way to tell an old expense from a burden still to come. Treating every attachment to a tool as waste would miss the real benefits of being skilled with it and of it working well with other things.
Read the economic working paper on switching costs.
A lab experiment: a warning sign can prompt a rethink instead of an automatic exit.
Marlone D. Henderson, Peter M. Gollwitzer and Gabriele Oettingen’s 2007 study had 183 German-speaking participants at the University of Hamburg. Each person chose a strategy for a computer task, presented as a general knowledge test, and gave a reason for it.
The researchers controlled what feedback people got on how they were doing. They also compared different plans for what to do when the feedback was disappointing. One plan was to think about how the strategy had been going. Another was to switch strategies.
When their scores had been falling, 70% of the group with the thinking plan switched, compared with 39% of the group with no plan. When their scores had been improving, the figures were 35% and 32%. The group with the switching plan switched at 58% after a fall and 57% after an improvement.
So people with the thinking plan switched more often when their recent scores had fallen than when they had risen. People with the switching plan switched at about the same rate either way. That’s why a sign that it’s time to review doesn’t have to decide the answer for you. It can start a comparison that still pays attention to what has been happening lately.
Why it matters: These were choices of strategy after feedback the researchers set up in a lab. The study didn’t show which choice led to better results with work tools.
Read the experiment on reflection and disengagement.
A small decision experiment: make the other uses of your resources visible.
Gregory B. Northcraft and Margaret A. Neale’s 1986 experiment involved 20 business students at the University of Arizona. They looked at two made-up building projects that had run into money trouble.
Each student saw two versions of the cases. One version left the students to work out other ways the remaining money could be used. The other spelled those uses out and asked students to think about them. Twelve students changed their decisions between the two versions. All 12 chose to sell when the other uses were spelled out, and to finish when they were only implied.
An opportunity cost is what you give up when you choose one use of your time or money over another. In this experiment, putting those other uses in plain view changed the decision for some people. Carrying on with the project was itself a choice about where future money would go.
The study was small and the cases were imagined. The authors also said that people might have been answering what the task seemed to want from them. And because the students were both given the information and asked to think about it, we can’t tell which of the two made the difference.
Why it matters: For a tool decision, the useful question runs both ways: what would moving to a new tool push aside, and what would staying with the old one stop you from doing?
Read the experiment on opportunity costs.
An official investigation: find out what actually went wrong before choosing a replacement.
The National Audit Office’s 2017 investigation looked at the WannaCry cyber attack on the NHS. It found that most of the infected NHS computers were running Windows 7, which was still supported but hadn’t been updated. Installing the update released in March 2017 would have protected them.
That finding matters because the cause changes what you should do about it. When an existing system fails, the problem can be upkeep that wasn’t done, rather than simply the system’s age. Here, the investigation pointed to one specific security update.
Why it matters: This was a recorded factor in a major incident, not a study of why people made the choices they did. It doesn’t mean updating fixes every problem with an older system. It does show why keeping, fixing and replacing each deserve their own look.
Read the official investigation of WannaCry and the NHS.
A report from regulators: moving the data isn’t the same as getting the work done.
A 2022 account from two UK regulators, the Bank of England’s Prudential Regulation Authority and the Financial Conduct Authority, describes how the bank TSB moved to a new system in April 2018. The data moved across successfully. But the new system failed straight away, and customers couldn’t use their banking services.
The regulators found weak control of the move and poor handling of the risks of relying on outside suppliers. The lesson is concrete: one part of the move worked, while the service people needed did not.
Why it matters: For a tool of your own, this makes being ready a separate question from whether your files will import. The real test is whether you can still do the essential work afterwards. The banking case is far bigger, and it doesn’t show that staying with the old system would have been better.
Read the regulatory account of TSB’s migration.
A product announcement: a firm date can help you prepare.
In July 2017, Adobe’s communications team announced plans to stop updating and offering Flash Player at the end of 2020. Adobe encouraged people to move to open formats.
That notice gave people a fixed date to plan their preparation around. A confirmed change in support tells you something quite different from a vague feeling that a tool is going out of fashion.
Why it matters: This is a record of a plan that was announced. It isn’t evidence about when people moved or why some waited. The useful lesson is to match a confirmed change that affects your work with the time a move would take.
Read the Adobe Flash retirement announcement.
What this means for you
Compare the work you need to do from now on. Count the benefits of knowing a tool well, but leave out the wish to make your past effort worth it.
If a system belongs to your employer, affects safety or is regulated, follow the process set by whoever is responsible for it. Never use this exercise to make changes to those systems that you aren’t allowed to make. If carrying on with one becomes unacceptable, use an approved backup or pause the work it affects until the problem is sorted.
Separate the old expense from the useful capability
List what the familiar tool helps you do now. It might let you finish tasks quickly, reuse templates, share files or work with other people. Check those benefits against real work rather than assuming that knowing a tool well proves them.
The time you spent making a template is already spent. But the future work that template saves still counts. So does any work you’d need to make it again somewhere else.
Use the same tasks, the same standard of quality and the same stretch of time for both options. Include future training, moving files and checking that the new setup works. Also count any time spent running both tools during the change, or working things out with other people. Keep hours, money and risks separate unless you have a clear reason to turn one into another.
Don’t count the same burden twice. If your estimate of how long a task takes already allows for working more slowly while you learn, don’t add that slowdown again as a separate cost. Training you do apart from those tasks is different work, and it belongs in the comparison.
Make the choice depend on future work
Write down what each option would push aside. Moving to a new tool might take time you’d otherwise give to an important project. Staying might leave you without something you need to be able to do. Name the actual work affected, so you can weigh it against the benefits you expect.
Say the rest of the move needs 20 hours of training, 8 hours of migration and 6 hours of validation. Migration means moving your files and setup; validation means checking that they work properly afterwards. That’s 34 hours in this made-up example. Assume both tools meet the same quality standards, and that the weekly savings leave out those separate moving tasks.
Over 26 weeks, saving 1 hour a week gives 26 hours back. After the 34-hour transition, you’re 8 hours behind. Saving 2 hours a week gives 52 hours back, leaving 18 hours ahead.
| Fictional weekly saving | Saving over 26 weeks | After 34 hours of transition | What it suggests, before other differences |
|---|---|---|---|
| 1 hour | 26 hours | −8 hours | Retain or repair if there is no other advantage. |
| 2 hours | 52 hours | +18 hours | A tested switch becomes a candidate. |
The second result doesn’t settle whether the new tool is ready. It shows how a believable change in the expected benefit can change the comparison. If you could really save anywhere from 1 to 2 hours a week, that range alone can change which option looks worth it. Trying real tasks, or taking a step you can undo, may help you decide.
Count only the work still ahead at your next review. Time you’ve already spent on a pilot—a limited trial—becomes a past expense too. Use what the pilot taught you, without treating a switch as a way to pay back the effort of running it.
Choose a review signal and keep the answer open
Name something you could see happen that would reopen the decision. It could be a confirmed notice that support is ending for your version, a failed attempt to share an essential file, or the same task failing again and again.
Write down what you would accept as evidence, and when you would review it. A rival tool announcing a new feature doesn’t, on its own, tell you whether your own work would improve.
Keep the review signal apart from the conditions for leaving. The signal says the choice needs your attention. The conditions for switching say what the other tool must let you do, and how much better it must be to make the move worth it.
If new evidence changes what you’re looking for, write down the old rule, the new information, the new rule and why you changed it. If you make an exception to a rule, write down when you made it and when it ends. These notes let you see later why your rules changed.
Check repair and replacement against the same essential work
Find out what is actually failing before you assume you need another tool. A fix or a change of settings may be enough to make staying work. Compare that option with carrying on as you are and with a possible replacement.
Before you commit to moving, try a task that is typical of your usual work. Check any files you need to move out and bring in, the finished result, and whether the people you work with can use it.
Decide how much time to allow for learning before you judge the new tool. Set a limit on how much effort you’ll spend on this trial before deciding what to do next.
Check that you can recover, too. Decide what you’d need to get back after a problem, or to return to the old setup. Check that you can do this while staying within what you’re allowed to change. A polished demo or a successful import doesn’t answer every question about daily work.
If the new tool fails something essential, don’t let its speed make up for it. If an important answer is unknown, name the cheapest useful check and a date to decide by. A limited trial can answer an important question without you needing to know everything first.
Review in time to act
Set a yearly reminder so the decision doesn’t get forgotten. Review sooner when a confirmed change or problem affects your work. If there’s a known end-of-support or contract date, count back from it using the time you expect a move to take.
At each review, update the future costs and benefits, check whether the new tool is ready, and choose to retain, repair, pilot or switch. Write down the reason, what is still uncertain, one next action and the next date.
Decide how much better the new tool must be once you allow for the work of moving. This is your margin. If a small change in your estimates would wipe out its lead, a narrow lead may not be enough.
Switch when a workable alternative passes the tests of essential work and recovery, and its future advantage is bigger than the remaining work of moving by your chosen margin. A hard requirement can instead make carrying on unacceptable. Then arrange an approved alternative or backup, or pause the work it affects.
How about your daily life?
Think of a familiar work tool you could realistically change. Is sharing a file you need getting harder? Are you looking at a replacement without having weighed its benefits against the work of learning it and moving to it?
You might choose a writing routine with templates you reuse, a spreadsheet process you share with other people, or a design tool whose support is due to change. Choose a real concern, rather than assuming every familiar tool needs replacing.
Write down situations you’d like to look at. One good idea is enough.
| A tool and the work I use it for | What I want to understand or improve |
|---|---|
| ____________________ | ____________________ |
| ____________________ | ____________________ |
My chosen situation: ____________________
Choose a tool you have permission to change, with time to test a realistic alternative. Take that situation into the exercise. Choosing it is a reason to look at the decision, not a decision to leave.
Write your switching agreement
Start with your chosen situation. If you came straight here, choose a work tool you know well, that you have permission to change, and that you have time to compare with a realistic alternative.
If a system belongs to your employer, affects safety or is regulated, follow the process set by whoever is responsible for it. Never use this exercise to make changes to those systems that you aren’t allowed to make. If carrying on is unacceptable, arrange an approved backup or pause the work it affects.
- Define the job and the requirements. Name the tool, the results you must produce, the people you work with and what your files need to do. Write down how well it works now on a typical task. Say which needs for quality, access or safety can’t be traded for speed.
Job, permission and essential requirements: ____________________ - Set a review signal. Choose a confirmed change or a failure you can see that affects that job. Say what evidence you will accept and the date by which you will review it. The signal opens the decision; it doesn’t order a move by itself.
If this happens, I will review by this date: ____________________ - Compare the future options. Include keeping the tool, a workable fix and a possible replacement. Use the same work, quality standards and stretch of time for each. Keep future hours, money and risks separate, and include what each choice would push aside.
Leave out past learning and fees you can’t get back. Include the value of being skilled with a tool, and future training, migration, validation, coordination and recovery work. Count each burden once. Mark your estimates, and note anything uncertain that could change the answer.
Comparison period, future advantages, burdens and uncertainties: ____________________ - Define what the candidate must demonstrate. Name a typical task, a file it must exchange, a finished result, use by the people you work with, and a test of recovering after a problem. Say how long you’ll allow for learning, the most effort you’ll put into the trial and a date to decide by. If evidence is missing, name a useful next check or someone responsible who can help.
Essential tests, limits and missing information: ____________________ - Choose the conditional outcome. Retain if the current tool is still the better choice and still acceptable. Repair if a workable fix makes it good enough. Pilot if an important unknown can be tested within your limits.
Switch when the tests of essential work and recovery pass, and the future advantage is bigger than the remaining work of moving by your chosen margin. If a hard requirement makes carrying on unacceptable, arrange an approved alternative or backup, or pause the work it affects.
Decision, supporting evidence and next action: ____________________ - Give the agreement a review date. Set a yearly reminder, and review sooner when a confirmed change affects your work or you need time to prepare a move. At each review, count only the costs and benefits still ahead. Write down any changed rules, the reasons, the dates and when any exceptions end.
Next review, earlier trigger and any exception expiry: ____________________
Ask Jing
Become a Jingius.
Your next step
You’ve finished when you have a decision that depends on stated conditions, evidence for it and a next date, or a clearly named gap to fill. Keeping a useful tool and getting ready to leave one can both come from the same comparison.
A note on the evidence
The experiments are about strategy choices in a lab and imagined investment decisions. What they found helps explain how people make decisions. The real cases show problems with upkeep, moving and timing in particular settings.
Those cases don’t show why any one person stayed with a tool. Keep your own work, and the limits you actually face, at the centre of the comparison.
How we prepared this module
This module brings together the economics of switching costs, decision experiments and recorded technology events, and links them to a written comparison of your future options. The made-up sums show how changing an estimate changes the decision. The exercise puts those ideas together for one tool you use.
Sources
- Joseph Farrell and Paul Klemperer. (May 2006). Coordination and Lock-In: Competition with Switching Costs and Network Effects. Nuffield College Economics Working Paper 2006-W7. Preliminary draft; §2.1, pp. 12–13.
- Marlone D. Henderson, Peter M. Gollwitzer and Gabriele Oettingen. (2007; online 30 November 2006). Implementation intentions and disengagement from a failing course of action. Journal of Behavioral Decision Making, 20(1), 81–102. Published article; Study 1 Method, pp. 85–88; Table 2, p. 90; Discussion, pp. 90–91.
- Gregory B. Northcraft and Margaret A. Neale. (June 1986). Opportunity costs and the framing of resource allocation decisions. Organizational Behavior and Human Decision Processes, 37(3), 348–356. Published article; Method, p. 351; Results, pp. 353–354; limitations, p. 355.
- National Audit Office. (27 October 2017). Investigation: WannaCry cyber attack and the NHS. HC 414, Session 2017–2019; linked PDF edition dated 25 April 2018. Official investigation; §2.5, p. 18.
- Bank of England / Prudential Regulation Authority and Financial Conduct Authority. (20 December 2022). TSB fined £48.65m for operational resilience failings. Official regulatory announcement; opening migration narrative and regulatory findings.
- Adobe Communications Team. (25 July 2017). Flash & the Future of Interactive Content. First-party announcement; paragraph beginning “Given this progress”.
