“Innovation strategy” documents have a recognisable shape: a vision statement, five pillars, and a roadmap that dissolves the first time a customer complains loudly. The reason is almost always the same — the strategy described where to arrive without describing how a decision would be made differently on Tuesday.
What follows is the operational version: five things a business can actually change, each one small enough to start this quarter.
1. Shorten the distance to a real user
The single strongest predictor of whether a company builds useful things is how many people sit between the person deciding and the person using.
Concrete versions: everyone who makes product decisions does a support shift each quarter. Sales call recordings are open to the whole team. Five users a month talk to someone who can change something.
None of this is a programme. It is a calendar entry, and it beats most research budgets.
2. Make the cost of trying something small
Companies do not lack ideas. They lack a cheap way to test one, so every idea has to be argued into a roadmap, which means only ideas with a confident advocate survive — and confidence is uncorrelated with being right.
| Cost of testing an idea | What gets built |
|---|---|
| A quarter and a business case | Only ideas with a senior sponsor |
| Two weeks, needs approval | Ideas that sound sensible |
| Two days, no approval | Ideas that turn out to be right |
3. Kill things on purpose
Most organisations have no mechanism for stopping. Features accumulate, each with a small maintenance cost and a handful of users who would complain, and gradually all capacity goes to keeping things alive.
The fix is a scheduled review — once or twice a year — where every feature has to justify its existence with usage data. Most will pass in seconds. The two or three that cannot are where your capacity went.
Deprecating well matters as much as deprecating: warn early, provide a path out, and be honest about why.
4. Automate the tasks with cheap verification
The highest-return automations in an ordinary business are boring: document extraction, transcription, categorisation, report assembly. They share one property — when the output is wrong, someone notices immediately.
The automations that go wrong share the opposite property. Anything where a mistake looks exactly like a success needs a human in the loop, and the checking usually costs as much as the doing. We put numbers on that split in AI tools for business.
5. Write decisions down
Not process documentation — decisions. What was decided, what the alternatives were, and what would make you revisit it. A paragraph each.
Two things follow. New people stop relitigating settled questions, which is a startling share of meeting time in any company over about thirty people. And when a decision turns out badly, you can see what you believed at the time — which is the only way an organisation learns anything rather than just having opinions about the past.
This is the cheapest item on the list and the one most consistently skipped.
What to ignore
- Innovation labs disconnected from the product. They produce demos and no shipped changes, and everyone learns that innovation happens elsewhere.
- Idea portals. Collecting ideas was never the constraint; testing them was.
- Benchmarking against large competitors. Their constraints are not yours, and copying their solution imports their problems.
- Multi-year transformation roadmaps. They outlive their sponsor. Quarterly results that compound do not.
Common questions
How do we do this without a budget?
Four of the five items cost nothing but attention — user contact, cheap tests, killing things, writing decisions down. Only the automation item costs money, and it is the one with the clearest payback calculation.
What if leadership wants a strategy document?
Write one, and make its content the five operational changes plus what each will show within a quarter. A document that can be checked is a different object from one that cannot.
How do we know it is working?
Time from idea to something a user touched. If that number falls, the rest follows. If it does not, nothing else you changed mattered.
Does company size change the answer?
Only the third and fifth items. A team of five already knows what it decided and has nothing to kill. Everything else applies from day one.
Leave a Reply