Insight budgets grow quietly. A tracker started for one product manager becomes permanent. A database subscription is renewed because nobody wants to be the person who cancelled it. An agency relationship drifts from project pricing to a retainer. Five years later the function costs twice what it did, and nobody can say which parts of the spend change any decision.
This post sets out six ways to cut that spend, drawn from insight cost reviews on both the client and agency side, and explains why starting from the decision is the one that matters most.
Start with the decision to be made
The single most useful question in an insight cost review is: which decision does this piece of work support, and who makes it?
Go through the budget line by line. For every study, tracker, subscription and report, name the decision and the decision-maker. If you cannot, that is a candidate for cutting. If the decision-maker cannot remember the last time they used it, that is a stronger candidate.
This exercise usually finds that 20 to 30 per cent of the budget supports no identifiable decision. It exists because it once did, or because a stakeholder likes receiving it.
Reframing the function around decisions rather than deliverables also changes what gets commissioned next. Fewer studies, tighter briefs, more work that ends in a choice.
Automate repeated work with AI
A large share of insight team time goes on work that is repeated every month or quarter: coding open-ended responses, summarising reports, monitoring competitors and news, building the same charts, updating the same deck.
Much of this can now be automated well enough to be useful. Open-end coding, first-pass summarisation of qualitative transcripts, competitor monitoring and standard reporting are the obvious candidates. The team reviews the output rather than producing it from scratch.
The saving is not only in hours. It is in the ability to do the same monitoring with a smaller permanent team, and to answer routine questions in a day rather than a fortnight.
Be realistic about what automation cannot do. It will not design a good study, run a sensible interview, or tell a management team what to do. Automate the repetition; keep the judgement.
Cut subscription costs
Data and syndicated research subscriptions are where the fastest savings usually sit. Typical findings:
- Two or three databases with overlapping coverage, each bought by a different team.
- Seats licensed for people who have left or never logged in.
- Subscriptions renewed automatically at list price when a negotiated rate was available.
- Reports bought annually that could be bought once every two or three years.
A subscription audit takes a couple of weeks. It lists every contract, its cost, its users, its renewal date and the decision it supports. Then it consolidates, renegotiates or cancels. Savings of 25 to 40 per cent on the subscription line are common.
Get better value from agencies
Agency costs are rarely the problem. Agency briefs are. A vague brief produces an expensive proposal with a broad scope, because the agency is covering the risk of not knowing what you want.
Three practical fixes:
- Write briefs that name the decision, the audience and the three questions that must be answered. Cut everything else.
- Separate design, fieldwork and analysis in quotes so you can see where the money goes and buy each part from the best-value source.
- Move from retainers to project pricing unless the retainer is genuinely used every month.
Fewer, better-briefed projects with two or three trusted suppliers usually cost less and deliver more than a long tail of agencies each doing one thing.
Stop research nobody uses
Trackers and recurring reports are the hardest to stop, because stopping them feels like losing information. Ask instead what would happen if the next wave did not arrive. If the answer is "nothing, for a quarter or two", reduce the frequency. If the answer is "nothing, ever", cancel it.
Duplicated reports are the other target. It is common to find brand health, customer satisfaction and NPS tracked in three separate studies by three separate teams, each measuring a slightly different version of the same thing.
Use a leaner team
Once repeated work is automated and unused research stopped, the permanent team can be smaller. The model that works is a small core of senior people who own the decisions and the relationships, supported by flexible capacity brought in for specific projects.
This is not about cutting headcount for its own sake. It is about matching a permanent cost to a permanent workload, and buying the rest when needed.
What a review delivers
An insight cost review produces:
- A clear list of possible savings, quantified and prioritised.
- An automation plan for repeated work.
- Subscription and supplier recommendations, including what to consolidate, renegotiate or cancel.
- A more flexible staffing model.
- Immediate actions and a 90-day plan.
A global FMCG company used this approach to reduce consumer insight costs by SEK 7 million over three years while maintaining the same level of delivery. The principle is simple: spend nothing on information nobody uses, and invest in supporting the decisions that matter.