Evidence Based Scoring gives each of your candidates a score showing how well they match the requirements you’ve set for the role, along with the evidence behind it. As soon as a candidate applies, Tali reads their CV and application form, lines up what it finds against your requirements, and records the passages from the candidate’s own words that support each match.

The score answers one question: how qualified is this candidate for this role? You can open up any part of a score to see what it was based on before you act on it.

Note: you will need access to the below permissions to work with Evidence Based Scoring:

Your System Administrator manages these against your role, and you can read more about how in our Managing Roles and Permissions article.

Four things are worth understanding before you start working with scores, because they shape what the score can and can’t tell you.

You can see the reasoning, not just the number. Every part of a score can be opened up and inspected. For each requirement you’ll see whether Tali found a match, how close that match was, and the exact words from the candidate’s CV or application form that it is based on. Nothing sits behind a number you have to take on trust.

The score informs the decision, you still make it. No candidate is progressed or rejected automatically on the strength of a score. Tali does the reading, and the hiring decision remains yours.

Nothing counts unless the candidate’s own words support it. Tali doesn’t credit a candidate with a skill because their job title suggests they probably have it, or because the market usually expects it in that role. If it can’t point at the passage that evidences a skill, it doesn’t claim the skill at all. That’s also why a claim with no supporting detail scores lower than one the candidate has genuinely described doing.

It’s more than a keyword match. A score doesn’t come from whether the right words appear somewhere in a candidate’s application. It takes into account what they have actually done, how long they’ve done it, how recently, how well evidenced it is, and how they work as well as what they know.

The score is calculated from the evidence found in a candidate’s application, using a set of factors: match strength, years of experience, recency and evidence strength. Each of those is explained below.

Today that means the candidate’s CV and their application form answers. Tali reads them and pulls out the skills, behaviours and qualifications they actually contain, along with the passage that supports each one. More source types, such as Data Capture forms and interview transcripts, are planned for later phases.

Two things are worth knowing about this step. First, it happens against the candidate rather than against a particular vacancy, so the job requirements can’t lead Tali into finding a skill that isn’t there. Second, it isn’t a one off. When new evidence arrives, such as an updated CV or an answer to a follow up question, it’s added to the candidate’s profile and the score is worked out again, so a score always reflects the strongest evidence available at that point.

Candidates are scored against the requirements set on your vacancy, not against a generic template for the job title. Those requirements fall into five groups:

Eligibility requirements are checked before anything is scored. If a candidate has confirmed they don’t hold something the role genuinely requires, they’re shown as Not Met rather than given a score that implies otherwise.

Once Tali has matched what the candidate has against what the role needs, four things decide how much each matched item is worth.

Not every factor applies to every requirement type. The explanations below describe how each factor works where it does apply; for the full breakdown of which factors apply to which requirement type, see our Evidence Based Scoring Best Practices article.

Match strength - How closely what the candidate has done lines up with what the role asks for.

Example: the role requires Team Leadership. A candidate who held the title “Team Leader” is an exact match. Someone who supervised a small team as a shift lead is adjacent, closely related with genuine transferability. Someone who once helped train a new starter is a weak match, loosely relevant and worth acknowledging but not much more.

Years of experience - Hands on time with a skill counts for more, but each extra year adds a little less than the one before. The step from no experience to a year of real practice is the biggest one, and by around ten years the difference between one more year and the last is small. So a candidate with three years against a skill scores well clear of one with a few months, while three years and five years sit much closer together.

Recency - Skills fade if they go unused, but not evenly, and how deeply someone held a skill decides how much of it survives a gap. Someone with several years of solid experience keeps far more of a skill over the same gap than someone who only picked it up briefly. Behaviours aren’t treated as fading this way, because a working style isn’t something you get rusty at.

Evidence strength - A claim backed by a real example carries more weight than one that is simply listed. Tali distinguishes between a skill the candidate has demonstrated by describing where and how they applied it, a claim only entry that is named with no supporting context, and something inferred that isn’t stated outright but is a reasonable read of the work described.

This is about whether the work was actually shown, not how impressive the wording sounds. A skill that appears only in a list of skills, in a run of keywords, under a heading, or as a self rating such as “Advanced” counts as a claim only, however senior or specialist the skill itself is, because all it tells you is that the word is on the page. To count as demonstrated, the candidate has to describe applying it somewhere: a project, a situation, or an outcome.

Evidence strength does both jobs: it’s one of the things that decides an item’s own result, and it’s also summarised on its own as your Evidenced signal, described next.

A candidate gets one headline score, and underneath it two sub scores you can read separately:

Reading the two together tells you more than the headline score alone. A high Capability with a low Evidenced score means a candidate who looks like a strong fit but hasn’t yet shown the evidence for it, which usually points to a follow up question rather than a lower ranking. Their score can climb once that evidence arrives.

A candidate saying nothing about a requirement and a candidate saying they don’t meet it are treated differently.

If a candidate’s application simply never mentions something, that’s recorded as unresolved and flagged as a gap to fill, not scored as a failure, so a candidate isn’t marked down for what their application didn’t happen to cover. That said, a required skill with no evidence at all can’t add anything to the score either, since there’s nothing to credit, which is exactly why it’s worth chasing as a genuine gap rather than something to leave unresolved. If the candidate has explicitly said they don’t have something, that’s recorded as a confirmed no.

The score is shown against the candidate as you work through your applicants, and the detail behind it lives on the Score tab in the candidate’s Application accordion.

Inside the Score tab, the candidate analysis is grouped the same way your requirements are: required skills, preferred skills, behaviours, preferred qualifications, and any other skills Tali found that the role didn’t ask for. You can switch between a detailed view and a compact view depending on whether you want to read the evidence or scan the shape of the match. The descriptions below describe the detailed view; the compact view shows the same information at a glance, with less of it expanded.

In the detailed view, each item shows its match strength, its evidence strength, the quote it’s based on and where that quote came from, along with the job title and company it was attributed to, how many years of experience it represents, and when it was last used. Where a skill couldn’t be tied back to a specific role, you’ll see that stated rather than guessed at.

Alongside the analysis you’ll find the eligibility requirements for the role and their status, and a set of recommended actions: the specific things that would raise this candidate’s score, and roughly how much difference each one could make. These are the follow ups worth making before you decide on a candidate, whether that’s confirming a licence or asking for an example of a skill they’ve only listed.

For a fuller set of best practices on setting requirements up so scores are as useful as possible, see our Best Practices article.

Set your requirements carefully - Candidates can only be matched against what the vacancy actually asks for, so the more precisely a role’s requirements are defined, the more useful the scores against it will be. Our Evidence Based Scoring Best Practices article walks through how to do that in detail.

Read the Evidenced score alongside the Capability score - A candidate who looks capable but is thinly evidenced is worth a question rather than a lower place in your list.

Fill the gaps before you decide - Where the score flags something unresolved, that’s a follow up or an interview question. Answering it raises the score if the evidence is there, and gives you a clear answer if it isn’t.

Use the evidence in your own conversations - The quotes and the reasoning are there to be used, whether you’re talking a shortlist through with a hiring manager or looking back at a decision months later.

Remember what the score is measuring - It describes one candidate against one role’s requirements, so it isn’t a ranking of one candidate against another.