Assessment in volume hiring, how to read the numbers in cases and award entries

Every autumn the UK trade publication Personnel Today publishes the shortlists for its HR awards. On 29 September 2026 it released the shortlist for the Innovation in Recruitment category, and it makes useful reading for anyone working with assessments in high-volume recruitment. Not because it settles what works, but because it shows how employers and vendors describe their own results when they want to persuade a panel of judges.

According to Personnel Today, the judges were looking for effective approaches to targeting, selecting and recruiting talent at all levels, from school leavers to senior positions. Several of the six shortlisted entries deal with large candidate pools, and three of them put some form of assessment at the heart of their solution.

The first is AWE, a UK defence and national security organisation, which together with the vendor Spotted Zebra introduced a selection platform with psychometric profiling and adaptive testing for early careers recruitment. According to the entry, the platform handles over 4,000 candidates a year. The organisation states that first-year attrition has fallen to below 10 percent, that offer rejections have halved, and that no adverse impact was found across demographic groups. AWE also reports savings of more than one million pounds.

The second is BeCareers, a Doncaster startup that has developed a CV-aware AI interviewer conducting conversational assessments. The company states that the tool can carry out 100 assessments in 24 hours, that screening time per candidate drops from up to 30 minutes to under five minutes, and that time-to-hire can be shortened by up to five days.

The third is Lloyds Banking Group, which with FourthWall shifted its early talent attraction from volume-driven engagement to what it calls structured behavioural alignment. The Faces of the Future campaign covered 51 events and produced 7,373 data captures, reported as an 89 percent increase on the previous year.

Figures like these deserve a sober reading. An award entry is written by the organisation that wants to win. The numbers have not been verified independently, and Personnel Today presents them as the entries' own claims. That does not make them wrong, but it makes them something other than evidence.

Take the attrition figure at AWE. A fall in first-year attrition to below 10 percent sounds impressive, but the entry does not say what the starting point was, how many hires the measurement covers, or whether other things changed at the same time, such as pay, onboarding or the labour market. Without a comparison group, the improvement cannot be attributed to the assessment tool alone.

The same applies to the claim of zero adverse impact. It depends entirely on how it was measured. Which groups were compared, at which stage of the process and with which statistical method? The entry does not say, and those are exactly the questions a buyer should put to the vendor.

The BeCareers figures are mainly about speed. One hundred assessments in a day and under five minutes per candidate say something about capacity, but nothing about whether the assessments are accurate or whether they rank candidates in a way that relates to later job performance. With AI-based interviews it also matters how the model was trained and how it documents its judgements, so that they can be explained to the candidate.

Cases from awards and vendor websites are useful as inspiration and as a starting point for questions. They are not a substitute for validity documentation. When choosing an assessment for volume hiring, you can use published cases to formulate concrete requirements.

Ask for the baseline. An improvement figure only means something if you know the starting level and the measurement period. Ask for the sample size. Results from a few hundred candidates tell you less than results from several thousand. Ask for the method behind any fairness claim. A vendor that can show how adverse impact was calculated has usually thought about it.

Ask as well whether the results were achieved with a client whose roles and candidates resemble yours. A solution that works for early careers in the defence industry will not necessarily have the same effect in retail or customer service, where the candidate profile and job requirements differ.

Finally, separate process gains from quality gains. Shorter screening time and fewer manual steps are real benefits and can be reason enough to change tools. But they are not the same as better hires. That requires the test results to relate to how people perform in the job, and that relationship has to be measured, not assumed.

The Personnel Today shortlist shows that assessment has moved into the core of how UK employers handle large candidate volumes. It also shows that the industry still tells its success stories in percentages and savings rather than validity data. The reader's job is to keep those two apart.

Kilde: https://www.personneltoday.com/hr/personnel-today-awards-shortlist-2026-innovation-in-recruitment-award/