Deciding when to cut a tool comes down to two questions: does a real workflow break without it, and would you buy it again today at full price knowing what you know now. If the answer to either is no, the tool goes, no matter how reasonable it felt the day you bought it.
- A tech stack rarely balloons from one bad call. It grows one reasonable purchase at a time until the total stops being reasonable.
- The keep rule is a single test: a real workflow breaks without this tool, and it is the only thing that fixes that break.
- The cut rule is a single test: you have not touched it in ninety days, or something you already pay for does the same job.
- Overlap is the quietest money leak in a stack because every overlapping tool felt justified on its own.
- The smoothest operations run fewer tools used fully, not more tools used partly.
- An audit works best on a schedule, not in a panic the week the annual charges hit.
There is a sentence I said out loud this year that a lot of owners will recognize. My third quarter blew my fourth quarter tech stack, subscriptions and all of that. So for me to stay under for the year, I have to go and cut shit down. That is not a failure of discipline. That is what a stack does when every single addition made sense in the moment. Knowing when to cut a tool is the skill that keeps the total honest.
Less is Best. I mean that as an operating rule, not a slogan. The businesses I see running the smoothest are almost never the ones with the longest list of logins. They run a tight set of tools, each one used all the way down, and they can tell you in one sentence why each one is still there.
What is tool sprawl, and when should you cut a tool?
Tool sprawl is the slow accumulation of software subscriptions that each looked reasonable at purchase but together drift into overlap, waste, and a total nobody signed off on as a whole. You cut a tool the moment it fails both halves of a simple test. The first half asks whether a real workflow breaks without it. The second asks whether something you already pay for could do the job. When a tool clears neither bar, it stays on the invoice out of habit, not value.
Nobody decides to end up with a dozen half-used tools. It happens the way weight happens. One subscription here because a project needed it. One there because a free trial quietly converted. A few because a tool was the shiny answer to a problem a process could have solved. Each one was defensible alone. The stack is what you get when you only ever evaluate the next purchase and never the whole.
A stack rarely grows from one bad decision. It grows from a hundred reasonable ones that nobody ever added up.
The keep rule, as one testable question
Here is the keep rule. A tool earns its place when a real workflow breaks without it, and it is the only thing that fixes that break. Both parts matter. Plenty of tools are pleasant. Pleasant is not the bar. The bar is: if this disappeared tomorrow, would something you actually do every week stop working, and is this the one thing that keeps it working.
Run the test honestly and most tools sort themselves fast. The ones that survive are the load-bearing ones. The payment flow. The thing your clients log into. The automation engine that moves work between systems without a human copying and pasting. Those stay because removing them breaks something real and nothing else covers it.
The cut rule, as one testable question
Here is the cut rule. A tool goes when you have not used it in ninety days, or when something you already pay for does the same job. Either condition is enough. You do not need both. The ninety-day silence tells you the tool was a want, not a need. The overlap tells you that you are paying twice for one outcome.
Overlap is the sneaky one. It hides because each overlapping tool was justified on its own day, in its own meeting, for its own reason. Put them side by side and you find three things that send reminders, two that store files, two that build pages. Nobody chose redundancy. It assembled itself. The guidance from application rationalization frameworks is to consolidate look-alike tools on purpose, which is just the cut rule with a longer name.
Why does a reasonable stack become an unreasonable bill?
A reasonable stack becomes an unreasonable bill because each tool is evaluated at the moment of purchase and almost never re-evaluated after, so the total grows unchecked while every individual line still looks fine. The fourth quarter is when this catches up with you. Annual renewals cluster. Charges you forgot about land in the same window. The math you avoided all year arrives at once, and suddenly the stack you built one calm decision at a time demands one uncomfortable reckoning.
The research names the costs cleanly. A bloated set of tools signals wasted resources, not efficiency, with hidden spend and overlapping functions that regular audits would catch. That is covered well in this piece on cutting tech complexity. The waste usually traces back to two gaps: nobody owns the tool, and nobody has the usage data. So renewals happen out of caution instead of evidence.
The two questions side by side
Here is the whole framework in one place. Run every tool through both rules. Most will give you a clear answer in under a minute.
| Rule | The question | What a “cut” answer sounds like |
|---|---|---|
| Keep rule | Does a real workflow break without this, and is it the only thing that fixes that break? | “Nothing stops working if I kill it.” or “Two other tools already do this.” |
| Cut rule | Have I used it in the last ninety days, or does something I already pay for do the same job? | “I haven’t logged in since spring.” or “This overlaps with what I already have.” |
The smoothest businesses are not the ones with the most tools. They are the ones that can explain every tool they keep in a single sentence.
How do you run the cut without breaking things?
You run the cut by listing every subscription first, tagging each one as load-bearing or optional, then applying the keep rule to the load-bearing set and the cut rule to everything else before you cancel a single thing. Order matters here. Cutting in a panic the week the charges hit is how you accidentally kill the one automation that was holding three processes together. The point is a calm pass, not a frantic one.
A practical sequence looks like this:
- List every recurring charge in one document. All of it, no exceptions.
- Mark each one load-bearing or optional based on whether a workflow breaks without it.
- Apply the cut rule to the optional pile first. Ninety-day silence and overlap are fast calls.
- Stress-test the load-bearing pile with the keep rule. Confirm each is the only thing fixing its break.
- Consolidate overlap on purpose. Pick one tool per job and move the work to it before you cancel the others.
The overlap step is where people get nervous, and they should move carefully there. Teams renew forgotten tools out of caution because they lack usage data, a pattern laid out well in this breakdown of unused software subscriptions. Caution is fine as a trigger to look. It is a terrible reason to keep paying.
What a tight stack feels like
A tight stack is quiet. You know what each tool does. You know who owns it. You know why it is still there. No fluffery, no logins you forgot you had, no tool running on bubblegum, popsicle sticks, and a prayer because it half-does a job something else could do fully. That quiet is the goal. The list getting shorter is just how you get there.
The term “application rationalization” comes from enterprise IT, where large organizations formally sort every piece of software into keep, replace, retire, or consolidate. The idea that you should periodically decide whether a tool still earns its place is not a startup trick. It is standard practice at companies with thousands of applications, scaled down to fit a business of one.
Almost every time I open a client’s list of subscriptions, I find at least three tools doing the same job and nobody can tell me which one is the real one. The pattern is never recklessness. It is that each tool was bought to solve a specific moment, and the moment passed, but the charge did not. The owner who said the fourth quarter line to me was not disorganized. She was just finally adding up a year of reasonable yeses. That addition is the whole skill. The keep rule and the cut rule exist so you never have to do it in a panic.
How do I know if I have too many tools?
You have too many tools if you cannot explain in one sentence why each one is still on the invoice. List every recurring charge in a single document. If any line makes you pause or guess at its purpose, that pause is your answer. Overlap and forgotten logins are the clearest signs the total has outgrown the need.
When should I cut a tool I pay for?
Cut a tool when you have not used it in ninety days or when something you already pay for does the same job. Either condition alone is enough. You do not need to prove both. The ninety-day silence means it was a want. The overlap means you are paying twice for one result.
What is the best way to audit my software subscriptions?
The best way is on a schedule, not during a billing panic. List every charge, tag each as load-bearing or optional, apply the cut rule to the optional pile, then stress-test the rest with the keep rule. Move overlapping work to one chosen tool before you cancel the others so nothing breaks mid-cut.
What is the difference between the keep rule and the cut rule?
The keep rule tests whether a tool is load-bearing: does a real workflow break without it, and is it the only fix. The cut rule tests whether a tool is dead weight: ninety days of silence or overlap with something you already pay for. One protects what matters. The other clears what does not.
Does a bigger tech stack mean a more efficient business?
No. A bigger stack usually signals wasted resources, not efficiency. Overlapping functions and hidden spend pile up when tools get bought but never re-evaluated. The smoothest operations run a smaller set of tools used fully. Efficiency comes from using each tool all the way down, not from owning more of them.
How often should I review my tools?
Review quarterly at minimum, with a deeper pass before the fourth quarter when annual renewals cluster. A regular rhythm keeps the total honest and stops the year-end reckoning from becoming a surprise. Short, frequent reviews are easier than one dreaded audit, and they catch overlap before it compounds into real waste.
Pick one afternoon this week and list every recurring charge in a single document. Run the optional ones through the cut rule and the load-bearing ones through the keep rule. You do not have to cancel anything that day. You just have to finally add it all up, because the fourth quarter bill is going to add it up for you either way.
If this framework would help someone on your team, send this to the person who keeps adding tools. You can also find more of how I think about tight operations over at the community.
… and see more of my posts in your Google results.
