Mission scoring: align profile and need without inventing useless criteria

The real problem: we send off-target profiles because we misread the need

The client asks for "a senior on cloud and industrialization." You send a very strong front-end profile with a different stack, because the CV mentioned "cloud" once in a project line from 2019. Nobody cheated. Sales skimmed the CV, the need was summed up in three words, and deadline pressure did the rest.

This scenario repeats every week at most ESNs. The outcome:

• The client gets a profile that doesn't fit and loses confidence in your ability to understand their need
• The candidate goes through an interview for nothing and ends up in an awkward position
• Sales burned a shot on a role where they could have sent the right profile on the first try

Scoring isn't a magic tool. It's a filter that forces you to pause for three minutes before hitting "send." Not to tick boxes, to check that what you read in the candidate's track record actually answers what the client described.

---

Why classic scoring grids fail

Most ESNs that tried scoring went through the same phase: an Excel sheet with fifteen columns, scores from 1 to 5, weights pulled out of thin air, and an overall score like "87/100" nobody knows how to interpret.

The syndrome of the spreadsheet that reassures internal teams

You build an exhaustive grid to show you have a process. In reality:

• Sales fill in the cells by gut feel so they don't block the send
• Scores vary by a factor of two depending on who evaluates, because there's no shared reference for what "3/5 in Java" means
• The client never sees this grid, it only exists to justify the send internally

Confusing eligibility and preference

A candidate can tick every technical box and be unavailable for two months. Another may have a gap on a specific framework but project experience perfectly aligned with the mission context. These are not the same decisions, and mixing them in the same column produces absurd rankings.

When everything is in the same soup, you no longer know whether a profile was rejected because a key skill was missing or because it scored 2/5 on a criterion nobody will look at again.

---

Build a matrix that actually works in the field

Useful scoring fits in four to six axes maximum, tied directly to the mission at hand. Not a generic model you apply to every RFP.

The axes that matter

Tenure in the exact domain of the need. Not "total years of experience." A developer with 8 years of experience but 6 months on cloud is not a senior cloud profile. Count years on the exact topic of the mission, and verify against mission descriptions, not the "skills" line on the CV.

Type of project environment. A mission at a large account with 40 developers on a legacy monolith and a mission at a five-person startup on a modern stack are not the same job. Someone who excels in one can sink in the other. Your scoring should capture:

• Usual team size
• Expected autonomy level
• Method framework (strict Scrum, flexible Kanban, firefighting mode…)
• Interaction with the business (direct or via a PO proxy)

Volume and criticality. A developer who has always worked on internal apps for 200 users will not react the same way to a production system at 50K requests per second. Same logic for regulated environments (banking, healthcare, defense): process rigor isn't a "nice to have," it's a cultural prerequisite.

Tools actually practiced. A tech list on a CV means nothing without context. "Docker" in a skills line can mean "I ran a container in training" or "I industrialized deployment for 30 microservices." Your scoring should cross the tool with the mission where it was used, duration, and level of responsibility.

What we leave out

• Vague "soft" skills like "team spirit," "rigor," "proactivity", impossible to score objectively; they clutter the grid
• Secondary techs the client listed "just in case" but that aren't in the real scope of the mission
• Certifications when the client hasn't made them an explicit prerequisite

---

Practical method: score in 15 minutes without kidding yourself

Here is a process that holds up day to day, tested with teams of 5 to 30 salespeople.

Step 1: Decode the client need (5 min)

Before looking at a single CV, restate the need in three blocks:

• The non-negotiable: without this, the profile is rejected in 30 seconds (main tech, availability, clearance)
• The core of the fit: what will make the candidate effective from the first sprint (similar context experience, autonomy, real seniority)
• The bonus: what differentiates your profile from the three others the competition is presenting

If you can't fill these three blocks, the client brief is fuzzy. Go back to them before scoring anything.

Step 2: Read the track record, not the CV (5 min)

The CV is marketing. The real track record is in mission descriptions:

• What did the candidate actually do on each mission? (not "participated in development," but "built the payment module independently")
• How long in each relevant context?
• What visible progression across missions?

A well-structured dossier lets you do this read in a few minutes. This is precisely where Betterfolio saves time: the track record is already organized by mission with techs, durations, and responsibilities clearly laid out.

Step 3: Fill the matrix and decide (5 min)

For each chosen axis, a three-level assessment is enough:

No score out of 100. No decimals. Sales must be able to look at their matrix and say in one sentence why they're sending this profile.

---

The five mistakes that kill credibility with the client

1. Inflated scores to force the send

Showing 92/100 on an average profile because the grid is poorly calibrated is worse than not scoring at all. The client who receives three profiles at 90+ and rejects two in the technical interview will stop trusting your evaluations. Ever.

2. Copy-paste across missions

When the grid is the same for every mission, answers become generic. A profile that's "good everywhere" on paper says nothing about fit with the specific need. Adapt your axes to each mission, even if the base stays stable, weights and non-negotiables change.

3. Scoring without knowing the client context

"The client wants React." Fine, but which React? Next.js SSR with performance tuning on a high-traffic app, or basic React for an internal back office? Without that level of detail, your technical scoring is a coin toss.

4. Ignoring the relationship track record with the client

A client who systematically rejects junior profiles won't change their mind because your grid says 78/100. Fold client feedback history into your calibration. If the last three "senior" profiles sent were judged too junior, your seniority dial is off.

5. Multiplying criteria to cover yourself

Fifteen criteria means zero criteria. Nobody reads a fifteen-row grid carefully. Every added criterion dilutes the signal from the important ones. The rule: if you can't explain in one sentence why this criterion is in the grid for this mission, remove it.

---

What changes for sales day to day

Faster client presentations

When your dossier is built on the same axes as your scoring, the pitch writes itself. You're no longer scrambling for what to say: you show three missions from the track record that each answer an axis of the need. The client reads the dossier, sees the logic, and the conversation moves from "can they do it?" to "how do we onboard them?"

Fewer profiles sent, more profiles retained

Under pressure, the natural reflex is to send five profiles hoping one sticks. Good scoring flips that: you send two profiles you can defend line by line. Conversion goes up, the client sees you as someone who gets their need, and the sales cycle shortens.

A common language between sales and recruiters

The recruiter sources on the same axes sales uses to score. No more "I thought that was enough" or "the client wanted something else." The matrix becomes the contract between the two: if the profile hits the axes, we send. Otherwise we explain why and adjust.

Actionable client feedback

When a profile is rejected, you can ask which axis it fell short on. The answer feeds straight into calibration for the next sends. Without a matrix, client feedback is often "it doesn't match", useless for improving.

---

Calibrating scoring over time: what almost nobody does

Most ESNs build a grid, use it for three months, then drop it because it "no longer reflects reality." The problem isn't the grid: it's the lack of calibration.

Monthly review of sends

Once a month, take the last ten profiles sent and sort them:

• Retained for interview: was scoring consistent? Were the axes right?
• Rejected before interview: which axis was missing? Was it in the matrix?
• Retained but not signed: the issue was elsewhere (availability, day rate, timing), scoring isn't the culprit

Adjust thresholds, not structure

If you find your "Strong" seniority profiles are regularly judged too junior by a client, don't add criteria, raise the bar for what counts as "Strong" for that client. The matrix stays simple, but its calibration sharpens.

Document edge cases

"Partial" profiles that still get through are the most interesting. They show what really matters to the client beyond the checklist. Note these cases: they become your best arguments for the next atypical profiles.

---

When scoring should step aside

Scoring cuts noise and structures the discussion. It doesn't replace judgment.

Situations where you should override the grid:

• The profile is "Weak" on a technical axis but project experience is so close to the client context that overall fit is obvious
• The client has a history of positive feedback on atypical profiles: they value adaptability more than the checklist
• The candidate has a fast, documented upskilling path that offsets raw tenure gaps

In these cases, sales must be able to send the profile by spelling out the gap: "This profile has only 2 years on Kubernetes but led the full migration of a 15-microservice IS at X, here's the context." That's defensible. An inflated score isn't.

Red flag: if scoring stops you from sending a profile you'd naturally defend in a client meeting, the grid has become an administrative obstacle. Reduce the number of criteria, clarify thresholds, and leave room for field judgment.

---

In summary: principles of scoring that hold up

• 4–6 axes max, derived directly from the mission need, not a generic template
• Three assessment levels (Strong / Partial / Weak), no scores out of 100
• Separate eligibility and preference: a yes/no is not a score
• Score the track record, not the CV: cross tech, context, and duration
• Calibrate monthly: compare scores with real client feedback
• Let judgment take over when the profile is defensible despite a weak axis
• A readable dossier so the client can verify fit themselves, without a methodology slide

Scoring isn't an end in itself. It's a translation tool between what the client asks for, what the candidate has done, and what sales can defend. When those three align on the same axes, conversion speaks for itself.