Methodology
How our matching works
Compare North doesn’t use one universal “best tools” list. Your answers determine what matters for your business. We compare those needs against researched provider data, remove options that cannot meet essential requirements, and calculate a personalised Match Score for the remaining choices.
Personalised
Different needs produce different rankings.
Deterministic
The same answers and same research produce the same result.
Independent
Commercial relationships do not influence scoring or ranking.
00The process
Your answers, our research and a fixed set of rules.
Every Finder on Compare North follows the same six stages. Nothing is generated at the moment you press submit — the questions, the research scores and the rules are all defined and reviewed in advance, then applied identically to every visitor.
01
Your answers
What you are building, what is essential, what you would prioritise.
02
Our research
Every provider scored 1–10 against the same criteria, region by region.
03
Eligibility
Providers that cannot meet an essential requirement are removed.
04
Weighting
Your answers decide how much each criterion counts for you.
05
Match Score
Weights multiplied by research scores, summed to a score out of 100.
06
Ranking
Ordered by Match Score, with fixed rules for ties and close results.
Stages run in this order. Eligibility always runs before scoring, so a provider that cannot do something you called essential is never scored, ranked or shown.
01Your answers
Your answers are the input, not a filter on a fixed list.
A Finder asks a short set of questions written in advance. Each answer option is mapped to one or more research criteria with a defined weight, so choosing an option is what makes a criterion matter for you.
Answers do different jobs. Some describe your situation and shape weighting; some are treated as essential requirements; one describes scope. Selecting more than one option in a question reinforces the shared criteria behind them.
Answer roles
What you are building
Establishes the purpose of the tool and the criteria that become relevant at all.
What is essential
Treated as a hard requirement. Providers that cannot meet it are removed, not penalised.
How much you need
A scope answer, used as a fixed part of the score rather than a weighted criterion.
What you would prioritise
Up to two priorities, which increase the weight of the criteria behind them.
Not sure
A valid answer. It adds no weight and never counts against any provider.
Website Builder Finder example
Selecting “Online store / ecommerce brand” gives weight to the physical-selling criteria and sets a minimum capability any recommended provider must reach. Selecting “Keep it simple and all-in-one” adds weight to ease of use and management.
Before any scoring happens, a Finder checks that you have given at least the minimum amount of usable information. If you have not, it says so instead of producing a ranking.
02Our research
Every provider is scored against the same criteria.
For each Finder we define a fixed list of criteria and assess every provider against all of them on a 1–10 scale. Scores describe capability, not popularity, and are recorded with the reasoning behind them.
Some criteria are context-sensitive. Value for money, for example, is assessed separately for general use, physical selling and digital selling, because the same pricing can be good value in one situation and poor value in another.
Scores can also be overridden per region where availability, pricing or payment support genuinely differ. Where a regional assessment exists it is used in place of the global one.
Research scale
- Best-in-class for this criterion9–10
- Strong and dependable7–8
- Workable, with limitations5–6
- Weak or awkward in practice3–4
- Not realistically suitable1–2
Website Builder Finder example
24 providers are assessed against 26 criteria, covering areas such as physical and digital selling, bookings, memberships, publishing, design quality, ease of use, technical flexibility and value for money.
A research score is never changed for a single visitor. Personalisation happens in the weighting, not in the underlying assessment.
03Eligibility
Some providers are removed before scoring begins.
Scoring alone can hide a dealbreaker: a very strong provider could still finish near the top while being unable to do the one thing you said you must have. Eligibility prevents that by removing providers before any score is calculated.
The checks run in a fixed order, and all essential requirements are tested — not just the first failure. A provider that fails a requirement cannot be recovered later by being a good fit in other ways.
Order of checks
Availability in your region
A provider that isn't properly available where you are is removed first.
Essential requirements
Every requirement you marked essential is checked against the research score for that capability. A provider must clear all of them — clearing most is not enough.
Scope fit
Finally, providers are checked against the scale of what you're building. Some are a clear fit, some are ruled out, and some are only suitable when specific conditions apply.
Website Builder Finder example
If you say advanced ecommerce features are essential, a provider must reach at least 6.0 / 10 on advanced ecommerce capability to remain in the running. Requirements that depend on more specialist capability, such as advanced integrations or complex scheduling, use a higher bar of 7.0 / 10.
If your combination of requirements leaves nothing suitable, the Finder tells you that directly. It will not fill the space with providers that don’t qualify.
04Weighting
Two people can get a different order from the same research.
Weighting is where personalisation happens. The research scores stay the same for everyone; what changes is how much each criterion counts towards your score.
You can mark up to two priorities. A priority doesn’t add points to a provider — it increases how much that criterion influences your score, which can move providers up or down.
Criteria you gave no signal for carry no weight at all. Nothing is assumed on your behalf.
How weight is built
Collect
Each answer you select adds weight to the criteria behind it. Answers that point at the same criterion add together.
Boost priorities
Criteria behind the priorities you chose are multiplied by 1.4 before anything is balanced out.
Balance to a whole
All weights are then rescaled so they add up to 100% of the weighted part of your score.
Reserve scope
When you answer the scope question, 15% of the score is reserved for scope fit and the weighted criteria share the remaining 85%.
Illustrative weighting
- Selling physical products40%
- Ease of use & management22%
- Affordability & value14%
- Design quality9%
Shape only, to show how weight is distributed. Your own weighting depends entirely on your answers.
05Match Score
A Match Score out of 100, built from your weighting.
Each criterion contributes its weight multiplied by the provider’s research score for that criterion. Add those contributions together, add the scope-fit part, and the result is a Match Score out of 100.
A Match Score is a measure of fit against your stated needs. It is not a rating, a review score, a probability or a prediction of how happy you will be.
The score is calculated at full precision and only rounded to a whole number for display, so two providers that look identical on screen may still be separated by a fraction of a point underneath.
The calculation
contribution = criterion weight × (research score × 10)
Match Score = sum of all contributions + scope-fit contribution
Research scores are held on a 1–10 scale and converted to a percentage of the available points for that criterion. When scope is answered, scope fit accounts for exactly 15 of the 100 points.
Worked example
Website Builder Finder
Beginner physical store
A first-time seller wants a simple shop for physical products.
Match Score
- Selling physical products54.2
- Scope fit15.0
- Affordability & value12.1
- Ease of use & management11.6
- Design quality & visual impact2.9
95.9 before display rounding
One worked example from the Website Builder Finder, shown to make the arithmetic visible. A different set of answers produces different weights and a different order.
06Ranking & results
Ranked by Match Score, with no discretionary edits.
Eligible providers are ordered by Match Score and the top three are shown. Nothing is promoted, pinned or reordered afterwards, and pricing or commercial information is only attached once the ranking is already decided.
When two providers score identically, the order is decided by a fixed sequence of tie-breaks rather than left to chance.
Each top result also explains itself: the reasons shown are drawn from the criteria that actually contributed most to your score, based on answers you gave.
Tie-breaks, in order
Higher Match Score at full precision.
Stronger performance on the priorities you selected.
Stronger performance on your strongest stated need.
Better scope fit.
A fixed, neutral ordering rule so the result never changes at random.
Honest result states
Not enough information
If your answers don't give enough to work with, no ranking is shown and the Finder says so.
No suitable options
If every provider fails an essential requirement, you get an empty result, not a substitute.
Fewer than three matches
You only ever see providers that qualified. The list isn't padded to fill the space.
Very close match
When second place is within half a point of first, it is labelled as a close match. The label never changes the order.
Changing your answers changes the ranking, because it changes the weighting. That is the method working as intended, not an inconsistency.
07Independence & limits
What this method can and cannot tell you.
A comparison is only useful if you can trust what shaped it. These are the commitments behind every Finder, and the honest limits of what a score can do.
Independence
Commercial relationships carry no weight
We may earn a commission when you sign up with a provider. That relationship is not an input to eligibility, scoring, weighting or ranking, and pricing and links are only attached after the order is final.
No provider can pay for placement
There is no paid position, no boosted listing and no manual override of a result.
No AI writes your ranking
Results are arithmetic over reviewed research data and predefined rules. Nothing about your recommendation is generated on the spot.
Research is versioned
Questions, criteria, scores and rules are published as a version, so the same version always behaves the same way, and changes are deliberate.
Limits
- A Match Score reflects the criteria we assess. It cannot capture everything about running a business.
- Providers change. Research is reviewed and updated, so a score reflects our assessment at the time of publishing.
- Short questions cannot describe every situation. A close second may suit you better for reasons we didn't ask about.
- Always check current pricing and terms with the provider before committing.
+ Technical detail
- Research scale
- 1–10 per criterion, converted to a percentage of that criterion's available points
- Priority selections
- Up to 2, each boosting its criteria by 1.4× before weights are rebalanced
- Scope-fit share
- A fixed 15% of the score when the scope question is answered
- Score precision
- Full precision throughout; rounded to a whole number only for display
- Close match
- Second place within 0.5 points of first, before rounding
- Results shown
- Up to 3 eligible providers
- Internal check
- Every score is recalculated a second way and must agree, or no result is returned
These values are part of the published configuration for the current version and apply to every visitor equally.
Try it
See the method applied to your own situation.
The Website Builder Finder takes about two minutes and shows why each recommendation fits your answers.