Short answer
If RICE, MoSCoW or weighted scoring still feels like guessing, you have diagnosed it correctly: it is guessing, formatted. Frameworks convert inputs into rankings and cannot create the inputs. What is missing is customer evidence, decision authority, or both. You do not need another framework.
The team adopted a framework last year. There is a spreadsheet. Every quarter it produces a ranked list, and every quarter the ranking gets overridden: by a large customer, by the CEO’s conviction, or by whatever engineering had already started before the meeting. The product managers privately admit that the Impact column is vibes with decimals.
Planning now takes longer than it did before the framework existed, and the decisions feel no safer. That combination, more process and no more confidence, is the signature of a scoring model running on inputs nobody trusts.
What is actually happening when the score feels hollow
A prioritisation score depends on three inputs, and most B2B SaaS companies between $2M and $15M ARR are short on all three at once. The framework is the only part of the system that is visibly working, which is exactly why it takes the blame. It is doing its job perfectly: converting whatever you feed it into a ranked list, with no ability to notice that the feed is guesswork.
Evidence: the column that is quietly fictional
Impact and Reach are claims about customers: how many accounts feel this problem, how sharply, and what they would pay or churn over. Those are answerable questions, and they are answered by talking to people. Without an interview cadence and win-loss data, the cells contain the team’s collective impression of a market, which is not the same object and is systematically biased toward whoever spoke most recently.
Strategy: the reason everything scores highly
A framework ranks candidates against goals. If the strategy has never explicitly excluded anything, every candidate serves the strategy, the scores compress into a narrow band, and the ranking becomes arbitrary at exactly the point where it should be decisive. Strategy is not the inspiring paragraph at the top of the deck. It is the list of things you have decided not to do.
Authority: the missing spine
Someone has to make the ranking hold when your largest account calls with a request that sits at number fourteen. A spreadsheet cannot do that. A person with standing and evidence can, and in most companies at this stage that person has not been named, which is why the override always wins and the team has quietly stopped believing in the process.
“Every founder who tells me their prioritisation framework is broken is describing a symptom. In every single case so far, the framework was fine and the company had no customer evidence, no written exclusions, or nobody empowered to hold a line. Swapping RICE for something else fixes none of those.”
The test: three questions before you buy another framework
Ask these before changing anything. They take about ten minutes and they are deliberately uncomfortable, because each one has an obvious right answer that most teams cannot produce. Ask them of two or three people separately rather than in a group, because disagreement between answers is itself the finding.
- Can your team cite specific customer conversations behind the top five Impact scores? Not dashboards, not tickets: conversations, with names and dates.
- Has the current strategy explicitly killed anything this quarter? Name it out loud.
- When the ranking was last overridden, was there a written reason, or just seniority?
Three noes means the framework was never the problem, and replacing it will cost a quarter and change nothing. One or two noes tells you precisely which of the three inputs to repair first.
What good looks like in 90 days
The target is not a better spreadsheet. It is a state where the ranking is boring because everyone already knows roughly what it will say, having heard the same customer problems described in customers’ own words. That state is reachable in a quarter, and the three components below are what it takes. None of them requires new tooling.
- Roughly 20–30 structured customer and prospect interviews in a tightly defined segment, feeding a problem ranking the whole team has heard first-hand.
- A one-page strategy that names what you will not do this year, visible to sales as well as product.
- A decision forum where an override requires evidence rather than volume, with the reason written down.
- One case, early and public, where evidence changed a priority that a senior person preferred. That is the moment the process becomes real.
Your options, honestly compared
There are three, and the middle one is genuinely viable more often than consultants admit. The deciding variable is not budget, it is whether the founder has the calendar space to run an interview programme personally and the appetite to hear things that contradict the current roadmap.
| Option | What it costs | What changes | Fails when |
|---|---|---|---|
| Do nothing | Nothing visible | Planning theatre continues; the real roadmap stays with whoever shouts loudest | Always, slowly |
| Founder-run | A day a week for a quarter | The founder runs the interviews and writes the exclusions page | The founder cannot protect the calendar, or cannot hear hard things |
| Senior help | A retainer, typically $5,000–$15,000 a month for fractional product leadership | Evidence base built and rankings made to stick by someone with standing | There is no internal team to execute the decisions |
Our own engagement is the third row: $8,000 per month for 20–25 hours, a 3-month minimum, with the first 60 days built around 60–80 customer and prospect interviews. That interview volume is not a flourish. It exists because this exact problem, a framework starved of inputs, is the most common thing we are called about, and it cannot be fixed from inside the building.
Not sure whether your problem is the framework or the inputs?
Thirty minutes, no pitch deck. You will leave knowing which of the three inputs is missing.
Book a Product Strategy SessionFrequently asked questions
Which prioritisation framework do you recommend instead?
Whichever one your team already knows. The framework is the last tenth of the problem and the evidence and authority are the other nine tenths, so switching models moves a small number and leaves the large one untouched. If you have no framework at all, start with the simplest thing that forces a ranking, and spend the energy you saved on customer conversations instead.
How many customer interviews are enough?
Patterns tend to stabilise around 20–30 conversations within a tightly defined segment, and roughly 10–15 focused interviews will often surface the dominant themes for a single narrow question. Below about fifteen you are collecting anecdotes rather than seeing a pattern. The number matters less than the discipline of doing them continuously rather than in one burst before a planning offsite.
Our PMs already talk to customers. Why is that not enough?
Support conversations and success check-ins are not discovery. They surface the opinions of your current users about your current product, which is a map of where you already are rather than of where you should go. Structured discovery deliberately includes prospects who did not buy and customers who churned, and those two groups hold most of the information you are missing.
Can we fix the authority problem without hiring anyone?
Sometimes. If the founder is willing to name a decision owner, sit in the forum, and visibly back an evidence-based call that goes against their own instinct, the problem is solved without a hire. In practice that reversal is the hard part, which is why an outside senior voice often works: the no comes from the process and the evidence rather than from someone telling the founder they are wrong.
Buying a new framework is the most available action and almost never the useful one. Fix the two inputs underneath it, and the ranking you already have will start telling you something you can act on. For the wider version of this problem, our guide to what a $25,000 to $50,000 budget delivers covers what an evidence programme costs to run properly.