Most money wasted on analytics is wasted at the point of purchase, before a line of code. Scope before price, prove it small, and judge a partner by their questions.

A fixed price for unscoped work is a red flag

Most money wasted on analytics is wasted at the point of purchase, before a line of code is written. An organisation decides it needs better reporting, picks a tool or asks for a price, and signs. A year later it owns a licence it half uses and a dashboard nobody trusts. The technology was rarely the problem. The buying was.

In practice, this is where our data strategy and leadership work comes in, and Why The AAP Is Not A Traditional Retainer covers useful related ground.

Three mistakes do most of the damage, and none of them shows up on the invoice as a mistake. They show up six months later as a stalled project and a budget nobody wants to discuss.

Buying a tool instead of an outcome

A platform produces nothing on its own. The licence is the cheapest, easiest part of the whole endeavour, which is exactly why it gets bought first: it feels like progress. It is not. The work that creates value is modelling the data, agreeing definitions, building trust, and changing how people decide. So do not sign off a tool purchase that is not attached to a named outcome and a first deliverable. If nobody can say which decision it will improve, it is too early to buy.

Buying a price before a scope

A fixed price quoted before any discovery is a guess with your name on it. Either the supplier has padded it to cover everything they cannot see, and you overpay, or they have underpriced to win the work and will recover the difference through change requests once you are committed. Both are bad for you. A firm number for unscoped work is not reassurance, it is a warning. Ask instead what they would need to understand before they could price it honestly. The quality of that answer tells you most of what you need to know.

A good scope is cheap to produce and saves a fortune. It is not a wish list of reports. It names the handful of decisions the work must support, in business terms. It states the source systems and an honest read on the state of that data, because that is where the real effort hides. It includes a first deliverable small enough to ship in weeks and useful on its own. It has a definition of done someone in the business, not just IT, would sign. And it says, explicitly, who will own and run the result.

Buying capacity you cannot use

Over-buying is the recurring, invisible waste. A large capacity feels safe, the salesperson will not talk you down, and the cost sits on the books producing nothing. Most mid-market organisations need a fraction of what they are sold. Buy for the workload you have today, agree in writing when you will review it, and make scaling up the deliberate decision, not scaling down a year later.

How to read a proposal, and what to ask

You can tell a great deal from the document itself. A good proposal is specific to you, referencing your systems and decisions, not generic enough to have been sent to anyone. It is staged into deliverables you could stop after, not one big lump with payment up front. It is honest about the state of your data. It is clear on ownership and handover. And it is clear on the running cost, not just the build, because a proposal silent on what it costs to keep alive is hiding half the bill.

You often learn more from the questions a partner asks you than from anything in their pitch, and you can turn them around. What would you need to understand before you could price this honestly. What is the smallest useful piece you could deliver first, and when. If we parted ways, who owns what you built, and can our own people run it. Where would you tell us not to spend money. What will this cost to run once it is live. A partner who names something you should not buy is usually worth more than one who says yes to everything.

So before you sign anything, write down the outcome you actually want, in one sentence, and a small slice you could prove first. Then judge every proposal against it. Scope before price, prove it small, design handover in: that is the buying advice we give clients before they spend a penny with us or anyone else, and the full version is in our How to Buy Data Analytics guide. It is also why our Analytics Acceleration Programme opens with a small paid proof rather than a guessed price, and why our analytics strategy work often starts by talking a client out of the largest version of what they asked for.

If any of this sounds familiar, talk to us about your data.

Related reading

• Stop Counting Dashboards. Count Decisions.
• What The Pub Trade Teaches About Your Numbers
• A Dashboard Nobody Opens Is Not A Tool Problem