Screening vendors price in incompatible units. One charges per user per month, another per job posted, a third per credit consumed, a fourth per hire. None of those is comparable to the others, and none of them maps to the thing driving your cost, which is how many applications you have to get through. Cost per shortlist is the unit that makes the comparison possible.
It is a simple calculation. The difficulty is that vendors have no incentive to help you make it, so you have to normalise their pricing yourself.
Why the Common Units Mislead
Per User Per Month
Prices your team, not your workload. A three-person team screening 4,000 applications a year pays the same as a three-person team screening 400. It is the most common model and the least informative, because your cost driver is invisible to it.
Per Job Posted
Prices your requisitions. A role attracting 30 applications costs the same as one attracting 600. It suits employers with steady low volume and punishes anyone with occasional high-volume campaigns.
Per Credit or Per CV
Closest to the actual driver, which makes it the easiest to model and the easiest to compare. The risk is unpredictability: a bad month of volume produces a bad month of cost, so ask what happens when you exhaust an allowance mid-campaign.
Per Hire
Superficially attractive because it ties cost to outcome. In practice it prices the wrong event, since the work is in the applications you rejected, and a role you screen but do not fill still consumed the capacity.
The Normalising Calculation
Convert everything to one number: annual cost divided by shortlists produced per year.
You need three figures. Your roles filled per year. Your average applications per role, taken from your own data rather than a benchmark. And the vendor's annual cost at that volume, which frequently requires asking them directly because published tiers rarely map to real usage.
A Worked Comparison
Assume 60 roles a year at 200 applications each, so 12,000 applications.
Vendor A charges per user per month for three users, giving a flat annual figure. Divide by 60 shortlists. Your cost per shortlist is fixed, and it falls if volume rises, which is the one genuine advantage of seat pricing.
Vendor B charges per credit where one credit screens one CV. Annual cost is 12,000 credits at whatever the rate is, divided by 60. Your cost per shortlist scales directly with applications, so it is predictable per campaign and variable per year.
Vendor C charges per job. Annual cost is 60 jobs at the rate, divided by 60, so cost per shortlist equals the per-job price regardless of application volume. Excellent if your roles attract 40 applications, expensive relative to value if they attract 600.
Now run the same arithmetic at double your volume and at half. The vendor that wins at your current volume frequently loses at the volume you expect in two years, and seat pricing versus credit pricing flips depending on which direction you move.
What to Include in the Denominator
A shortlist is only comparable if it means the same thing. Define it before you compare: a ranked set of candidates, against your published criteria, with a per-candidate rationale, ready to hand to a hiring manager.
If one vendor produces that and another produces an unexplained score you then have to justify yourself, you are not comparing like with like. The second output requires manual work you should add to its cost.
The Costs Vendors Leave Out
Four lines that belong in the comparison and rarely appear in a quote.
- Implementation and configuration. Ask for it as a number and ask a reference customer of your size how long it actually took.
- Manual work either side. Time spent preparing files for upload, or cleaning output before it is presentable, is real cost. Measure it during a trial.
- Overage behaviour. What happens when you exhaust an allowance mid-campaign. Blocked, throttled, or charged at a premium rate.
- Exit cost. Whether you can export your data and your decision records in a usable form. This matters for subject access requests as much as for switching.
Ceilings Matter More Than Price
A cheap tool that cannot handle your largest campaign is not cheap. Ask every vendor for the numbers rather than the marketing claim: maximum documents per batch, maximum files per upload, maximum file size, concurrent batch limit, and rate limits.
Ours are 500 CVs per screening batch, 50 files per upload request, 10MB per file, one concurrent batch per user. Ample for the overwhelming majority of UK campaigns and a genuine constraint if you run single batches above 500. Every vendor has a ceiling. The ones that will not state it are the concern.
Testing the Model Before You Buy
Run one role you have already filled through each shortlisted vendor, using the applications you actually received, and record four things.
- Total elapsed time including your own manual work either side.
- Whether the person you hired appears in the shortlist, and if not, why.
- Whether you would be comfortable showing a candidate the rationale for their ranking.
- The actual consumption, so you can check the vendor's volume assumptions against reality.
That last point catches a common surprise. Credit consumption is sometimes higher than expected because re-runs, failed parses and duplicate submissions all consume capacity.
When the Model Says Do Not Buy
If you fill twelve roles a year at forty applications each, your annual sift is a few hours of work and no tool will pay for itself on labour. The honest answer in that situation is that the case has to rest on consistency and defensibility instead, and if you do not need those, do not buy anything.
We would rather you reached that conclusion from your own arithmetic than discovered it after a year of licence fees.
Frequently Asked Questions
Which Pricing Model Is Best?
Credit or per-CV pricing is easiest to compare and tracks your actual driver. Seat pricing rewards high volume and penalises low. Neither is inherently better; what matters is which matches your shape.
Should We Negotiate on Unit Price or Volume?
Usually on the volume band and the overage rate, because those determine what you pay in the months that hurt. A lower headline unit price with punitive overage is often worse than the reverse.
How Do We Compare a Free Tier Fairly?
Model the point at which it stops working, then price the tier above it. Free tiers are evaluation tools rather than plans, and the limit is usually monthly volume or batch size rather than features.
Does Cost Per Shortlist Capture Quality?
No, and it should not pretend to. Use it to compare cost, then use the campaign test above to compare quality. A cheap shortlist you cannot defend is not good value.
What If a Vendor Will Not Quote at Our Volume?
Treat reluctance as information. A vendor unwilling to state cost at your stated volume is unlikely to become more transparent after you sign.
The Single Takeaway
Convert every quote to annual cost divided by shortlists produced, at your current volume and at double it. Define what a shortlist must contain so the comparison is fair. Add implementation, manual handling and overage. Then confirm with one real campaign.
That produces a number you can defend to finance, which is more than a licence fee comparison ever does.
This article sets out a method rather than benchmark figures. Obtain current pricing directly from each vendor.
Sources
Greenhouse. Greenhouse Pricing.
https://www.greenhouse.com/pricing
Workable. Workable Pricing.
https://www.workable.com/pricing
JobAdder. JobAdder Pricing.
https://www.jobadder.com/pricing
Information Commissioner's Office. Employment Practices: Recruitment and Selection.
https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/employment/recruitment-and-selection/
InsightCircle
