Sage says it has supported British businesses for more than forty years. Its plans do the whole job, and it has just put its AI assistant into every Sage Accounting plan. What they lack is a reason to stay and pay more that Xero or FreeAgent does not also give. The function that would give it has to be built by the group’s product organisation, which is changing its head.
This case is built from public sources, without access to Sage. Its goal is a working hypothesis, not a goal the company has stated, and it states no budget: none is public for the UK business. It replaces an earlier version, which read the whole group as a Stagnant Leader (A4), on a scale the method has since retired.
The situation. A mature market for small-business accounting software. Sage’s users already run their books on it; Sage bundles its AI assistant, payroll and HR into the plans and raises their price.
The goal. Raise what each existing customer pays each month by six per cent against the year to September 2025: three per cent by the end of September 2026, the full rise tested on the twelve months to May 2030. It is a working hypothesis of this case.
The posture. Niche Expert: grow what each customer pays, with an offer they cannot get elsewhere. The name describes the play, not a forty-year incumbent.
What blocks the goal. Nothing in the plans is a reason to stay and pay more that a rival does not also give. The promise of trusted AI runs ahead of its proof, and Sage prices the same plans two ways.
The decisions. Four, in the first six weeks, each starting with a test: have one function built that only Sage offers; give existing users their own price path and reward accountants on value; show the proof of trusted AI where users work; ask users what they want.
What is not known. Sage publishes no marketing budget for the UK business, and no revenue or customer count for these two products. This case states no amount, and it tests before it commits.
Plans that do the whole job, and no reason of their own to pay more
Sage in the United Kingdom is the UK business of The Sage Group plc. Its UK company reported £413.8 million of revenue for the year to September 2024, of which £371.1 million from software subscriptions, all products together, intercompany income included. The offer of this case is Sage Accounting and Sage 50. On 30 September 2025, Sage put its AI assistant, Sage Copilot, with payroll, HR and carbon accounting, into every Sage Accounting plan. Its brand line, “Authentic Intelligence”, heads its UK homepage.
The intuitive reading is that Sage is too dear. For a new customer its mid-tier plan is indeed the dearest of the three compared, by two to six pounds a month. The evidence says the level of price is not the problem. What hurts is that the same plans are sold to new customers at 90% off for 10 months while existing users see their prices rise, and that Xero and FreeAgent sell payroll too, Xero AI as well.
Software, sold by subscription. Sage creates value through software and captures it through subscriptions: about nine tenths of its UK company’s revenue, intercompany income included. It prices the content of the plans, and the functions come from the group’s product organisation.
Why it matters: what the plans contain is the strategy, and the function that is missing is not built in marketing.
A posture chosen by the situation, not by the history of the company
The method reads three facts. The market is mature: Making Tax Digital adds users and obligations, about 2.9 million individuals between April 2026 and April 2028, to a category that already exists. What customers buy is a product: a monthly plan of software features, with no long contract. And the lever that fits is to grow what each customer pays: nothing shows high churn, and Sage is already raising prices with its AI assistant.
M3 (growth curve) × M4 (economic value) × the lever chosen at Step 2. The matching row of the selection matrix reads: “Raising what each customer spends (ARPU) with add-ons: the usual play for specialists.”
The Niche Expert
A posture is the stance a situation calls for, not an identity. This one asks for an offer that sets you apart for a defined group of customers, and proof of it. Its two Fatal Brakes are the offer itself (Features) and the claim to it (Positioning). Here neither is failing, and neither yet does what the goal needs.
For the UK customers of Sage Accounting and Sage 50 at 1 October 2025 who remain customers, raise the average revenue per customer per month by six per cent against the year to September 2025. The roadmap runs for four and a half years, so the goal is staged: three per cent over the year to September 2026, six per cent tested on the twelve months to May 2030, both against the year to September 2025.
Its current value is not public: Sage can read it in its own billing data. The size of the goal sits just above the group’s weighted price increase of 5.5% in its last financial year.
Nothing works against the goal, nothing carries it
A8 has eight priority dimensions. Each score answers one statement: this dimension is helping Sage reach the goal, from −3 (completely disagree) to +3 (completely agree). A dimension scored on several questions can land on 0, when what helps and what hurts cancel out. The score does not rate the dimension in itself.
Features and Pricing are also the Growth Drivers of this posture. Each carries one score.
Five conditions help a little, three cancel out, and none works against the goal. None of the four that the posture asks most of reaches its level. Left as it is, a rise in what each customer pays rests on price and bundling, which rivals can answer.
Marketing holds most levers; the one it does not sets the calendar
Sage has no function that rivals lack and that its users value enough to pay more for: its answer has been to bundle what rivals also sell. It publishes no research that asks its users what they want. Its promise of trusted AI is not yet delivered in the software they run: its AI Trust Label went first to its product for mid-sized businesses. And one price system pulls partly the other way, for customers and for their accountants.
Marketing holds most of the levers: the survey, the evidence behind each claim, the messages, the price path, the rewards for accountants. Three decisions sit with other teams. Technology must date the label; Finance must approve prices and discounts; and the group’s product organisation must build the missing function, whose chief product officer steps down at the end of March 2026.
Build: a function only Sage offers, and a survey of its own users. Strengthen: the proof of trusted AI, and of each claim made for the assistant. Rework: the two price signals, and the rewards for accountants. Features and Positioning carry the same score and call for different work.
Four decisions, each starting with a six-week test
None carries an amount: no budget is public. Each one says what would change the answer.
The function only Sage offers
Commit the product organisation to build one function existing users would pay more for and rivals do not offer, or keep raising what each customer pays through bundling and price. A sponsor and a build date are asked for in week 1. What would change the answer: no function is named by users as worth paying for, or rivals already offer it.
Two price paths, and the reward for accountants
Give existing users a price path of their own and reward accountants for the value of their clients, or keep raising prices for those who stay beside deep discounts for those who arrive. Finance decides after the tests. What would change the answer: the requests for discounts do not come from users who saw the new-customer offer; practices with larger discounts do not add clients on cheaper plans.
Proof of trusted AI where users work
Carry the proof of trusted AI into Sage Accounting and Sage 50, and publish the method behind each claim made for the assistant, or keep the promise as a brand campaign. Technology is asked for a date in week 1. What would change the answer: the proof does not raise trust or credibility in the tests.
Knowing what users want
Survey existing users on what they aspire to and cite it in each brief to them, or keep relying on surveys of small businesses at large. First questions read by mid-January 2026; the survey in the field by August 2026. What would change the answer: the answers show nothing beyond productivity and AI.
The cost of not deciding the first one: the rise rests on price and bundling alone, which rivals can answer, while Xero moves a Sage customer’s books across for free.
Four and a half years, not a one-year price rise
Read before committing
Cycle 1 · Fix
Cycle 2 · Align
Cycle 3 · Scale
Cycle 1 lasts eighteen months, on Sage’s own precedent, and the next two take the same length: its AI assistant took about seventeen months from its group launch to every Sage Accounting plan, and about ten in the plans themselves; neither span counts finding and building a new function. It is one of the least certain estimates of the case.
Not the level of price, and not a defence of the base
The intuitive answer to a dearer plan is a lower price. The method reads the same facts differently.
- Sage is the dearest plan for a new customer
- Answer with deeper discounts
- The gap between the two prices widens
- Existing users may keep asking for the offer they cannot get
- The plans stay ones that rivals can answer
- Dearer by two to six pounds; the comparison flips with the plan chosen
- Sage 50 users renew at higher prices, on UK & Ireland figures that also cover Sage 200
- What hurts is two prices on the same plan, in full view of both
- What is lacking is a reason to pay more that rivals do not give
- What comes first is a function only Sage offers, and a price path for those who stay
Where this diagnosis can be contested. The earlier version of this case read the purchase as a service and Sage as a Stagnant Leader (A4), defending its base. The analysis reads it as software features sold month by month, while a rival converts Sage accounts for free: a defence of the base would answer a churn that no source shows to be high. If the market were read as declining, the same goal would call for Value Harvester (A6), maximising cash while investing as little as possible. Both calls are judgments.
Five lessons that travel beyond software
A complete offer is not a reason to pay more
Sage’s plans do the whole job. Its rivals’ plans do it too. Adding what they also sell raises the bill, not the reason to pay it.
Put the proof where customers work
A promise of trusted AI needs its proof in the software users run, not first in another product. Until then the promise stays a claim.
Two prices on one plan send two messages
Rises for those who stay and deep discounts for those who arrive, in view of both, work against a goal set on those who stay. So do rewards for volume.
Name the team that holds the slowest lever
Marketing holds most of the work here. The function that sets the calendar is built elsewhere: the first decision is a sponsor and a date from that team.
Say what a staged goal can no longer measure
The roadmap runs for four and a half years. Over that span, price rises alone may cover the goal: say so. A yearly rate held to the end would test more.
A Niche Expert is not always a software group with forty years behind it. The same posture in a sourdough bakery, a fictional example built for small businesses: Boulangerie Ferment →
Preparing next year’s plan and budget? Start from that situation: We are preparing the plan →
Is there a reason to pay you more that your rivals do not also give?
A complete offer and a clear promise do not by themselves give customers a reason to pay more. Find out in about 8 minutes which condition blocks your own goal.
A8 reference & full Vital 8 logic → marketingcanvas.net · The method in full → the book