Planning assumptions and where they should live

An assumption is any input to the plan that is not derived from another input and is not a fact. Pay awards, price uplifts, inflation rates, exchange rates, attach rates, churn rates, conversion rates and instrument life are all assumptions. They are the joints of the model, and they are where plans are most often quietly broken.

The governing rule is: one assumption, one place, one owner, one stated sensitivity. An assumption that appears in three models will diverge in three directions. An assumption without a named owner is owned by whoever was asked first. An assumption without a sensitivity cannot be challenged, because nobody knows whether it is worth arguing about.

The practical mechanism is an assumption register: a single document, versioned and dated, that lists every assumption, its value, the person accountable for it, the evidence behind it, and what a defined movement does to the result. It is issued before submissions are requested, not after. This is not bureaucracy. It is the thing that removes an entire round of iteration, because it stops business units debating whose inflation rate to use and forces them to debate the only things they can actually influence.

Three failure patterns are worth recognising.

The assumption embedded in a formula. A 3.5 per cent uplift typed directly into a cell rather than referenced from a register. It cannot be found, cannot be changed globally, and will still be there two plans later.

The assumption that is really a decision. "We assume a 2 per cent price uplift" is not an assumption; it is a commercial decision that someone must take and communicate to customers. Labelling decisions as assumptions is how organisations plan for outcomes nobody has agreed to deliver.

The assumption nobody owns. Pentland's instrument life of eight years is the clearest case. It drives retirements, which drive the installed base, which drives consumables and service, which together are GBP 28m of revenue at margins well above the instrument line. If it is really seven years, retirements are 486 rather than 425 a year and the installed base grows more slowly than planned every year, compounding. Nobody at Pentland owns that number.

Worked example 5: The assumption register, and what each line is worth

This is the register FP&A should issue in mid-September, before any submission is requested.

The working behind the first line, since it is the one candidates are most often asked to reproduce. Average salary across 540 staff is GBP 42,000, so the salary base is 540 x 42,000 = GBP 22,680,000. One percentage point is GBP 226,800. Employment on-costs, being social security and pension, add approximately 18.8 per cent, so 226,800 x 1.188 = GBP 269,438, call it GBP 270,000. That is 270,000 / 7,500,000 = 3.6 per cent of operating profit, and at Pentland today it is settled in a corridor conversation in November by whoever is asked first.

The other workings, briefly: a 1 point price change on planned instrument revenue of GBP 62.0m is GBP 620,000, which falls straight to gross profit. A 2 per cent consumption rate change is 3,568 x 5,729 x 0.02 = GBP 408,821 of revenue, and at 58 per cent that is GBP 237,116. Three points of attach on 760 new units is 22.8 contracts at GBP 4,397 = GBP 100,252 of revenue, GBP 66,166 at 66 per cent. Two points of churn on the renewable base of 1,844 is 36.9 contracts, GBP 162,000 of revenue, GBP 107,000 of gross profit. One month less distributor stock cover is 11,000,000 / 12 = GBP 916,667 of shipments, GBP 376,000 at the 41 per cent instrument margin. And 5 per cent on GBP 13.5m of Middle East revenue is GBP 675,000, which falls almost entirely to profit because the cost base is in sterling.

Now add up the adverse case. If the six largest assumptions each move one notch against the plan: 270,000 + 620,000 + 237,000 + 107,000 + 376,000 + 675,000 = GBP 2,285,000. Against an operating profit target of GBP 9.3m, that is 24.6 per cent of the profit line sitting in eight lines of assumption, none of which appears in the board pack today.

That number is the argument for the register. It converts "we should document our assumptions" from a hygiene request into a governance one. A board that approves a GBP 9.3m profit target without seeing a GBP 2.3m assumption range has not been told what it approved.