Four metrics that should replace total applications in your hiring report

Total applications is the most widely reported metric in early-careers hiring in Vietnam, yet it says nothing about which channel delivers qualified candidates and which merely delivers volume. VCO Group proposes four replacements, all measurable from data a programme already holds.

Why total applications is the most misleading number in the report

In most end-of-season reviews, the first figure on the page is the total number of applications received. It is easy to measure, easy to compare with last year, and easy to present to the board. The problem is that it collapses three completely different things into one number.

First, it conflates applications with candidates. In VCO Group's dataset the average number of applications per candidate is above one, because a share of candidates apply to several programmes in the same window. Reporting by application therefore always overstates the number of real people.

Second, it conflates qualified and unqualified candidates. A channel that delivers twice the volume at half the quality contributes no more than the other channel, while potentially costing twice as much.

Third, it conflates what the programme controls with what it does not. Application volume is affected by seasonality, by university exam calendars, and by whether a competitor opened a programme at the same time. None of that reflects operational quality.

Four replacement metrics and how to calculate each

All four metrics below can be derived from data a programme already holds, with no new measurement tooling.

Four metrics to report instead of total applications

The first three can be measured during the current season. The fourth requires joining marketing spend by channel with candidate data, and that is normally something to set up at the start of the season rather than at the wrap-up.

Why the median score beats the pass rate when comparing channels

Of the four, the second is the one most often substituted incorrectly. The natural reflex when comparing two channels is to calculate each channel's pass rate. That approach carries a systematic flaw.

The pass rate depends on the threshold the programme sets. If the threshold changes between seasons, or if two programmes within the same company set different thresholds, the rates are not comparable even when source quality is identical.

Median scores and percentiles do not have that problem. They describe the ability distribution of candidates from each channel directly, independent of any administrative decision about the bar. This is why every comparison in VCO Group reports is based on score distributions rather than pass rates.

There is a useful side benefit: once you report on score distributions, you can compare your results across multiple seasons even after changing the threshold, as long as the question bank and scoring conversion stay the same.

One metric worth adding and one worth dropping

Beyond the four main metrics, there is one cheap addition most programmes do not record: the share of candidates who are simultaneously in another company's programme.

This is captured with a single optional question on the application form. It is not used to screen anyone out; it tells you the real level of competition you face and helps calibrate expectations for offer acceptance in later rounds. In VCO Group data, roughly one in six candidates applies to more than one programme in the same window - and that is a floor, because the dataset does not cover the whole market.

Conversely, one metric should come out of the report, or at least off the front page: reach and impressions on recruitment marketing posts. It has almost no relationship with qualified candidate volume, it can be inflated with budget, and it creates a sense that the programme is performing well while the funnel behind it may be leaking badly.

The general rule for choosing what to report: keep only the metrics where a change tells the team exactly what to do next.

External benchmarks for reading your own numbers

A set of metrics is only useful with something to compare against. A few international reference points help in reading internal figures.

On application completion rates, CareerPlug's analysis of more than 10 million applications puts the cross-industry average at around 6 per cent, with very wide variation by sector. That is far lower than most people assume and shows that front-of-funnel drop-off is a general phenomenon, not one programme's failing.

On how to break metrics down, widely recommended practice is to split completion rates by device, by source channel and by seniority rather than reporting a single blended figure. A large gap between mobile and desktop points to an interface problem; a gap between source channels is a basis for reallocating budget.

For Vietnamese market context, TopCV's annual recruitment market report - built on a survey of more than 3,000 employer and worker representatives plus analysis of nearly 300,000 job postings - is a useful public reference when placing programme results against overall supply and demand.

Three things to do

1. De-duplicate candidates before counting, and report people rather than submissions.
2. Move every comparison between channels and between seasons onto median and percentile test scores instead of pass rates.
3. Set up the join between channel spend and candidate data at the start of the season, because cost per qualified candidate cannot be reconstructed at the wrap-up.

Frequently asked questions

Which metrics should replace total applications?

The four replacements are distinct candidates after de-duplication, median test score split by channel, test participation rate split by channel, and cost per candidate who passes the test round. All four are derivable from data a programme already holds.

Why should channels not be compared using pass rates?

Because the pass rate depends on a threshold the programme sets itself. If the threshold changes between seasons or differs between programmes, the rates are not comparable even on identical source quality. Median scores and percentiles describe the ability distribution directly and avoid the problem.

How many candidates apply to several programmes at once?

In VCO Group data, roughly one in six candidates applies to more than one programme in the same window. This is a floor, because the dataset does not cover the entire market, so the true number of programmes a candidate applies to in a season is higher.

About VCO Group

VCO Group operates the assessment infrastructure behind Young Talent Programmes and Management Trainee schemes in Vietnam, covering application intake, aptitude testing, automated interviewing and Assessment Centres. VCO Group serves young talent and candidates with under three years of experience, working with large enterprises and multinationals in Vietnam. Its assessment and candidate management platform is called Testalents. Head office in Ho Chi Minh City.

If you would like to benchmark your own selection data against the market, or discuss the Young Talent Program Insight Report, contact Mr. Dung on 0334 889 068 or visit vcogroup.com.vn.