Stop measuring onboarding by completion

Back in 2024, I was looking at onboarding in a growing sales business that had invested quite heavily in improving the experience for new starters. They had a proper programme, a learning platform, welcome sessions, product training, systems training, assessments and a reasonably impressive collection of things for people to complete during their first few weeks.

By most of the measures they were using, it worked. Completion was high, feedback was good and people generally seemed to enjoy the experience. The problem was that the Sales Director was still frustrated. New hires were taking far too long to become productive, managers were spending too much time correcting basic mistakes, and there was a sizeable gap between somebody having technically completed onboarding and anybody being comfortable putting them in front of a client without significant supervision.

That gap is one of the more expensive things growing businesses routinely fail to measure.

We tend to treat onboarding as an event with a beginning and an end. Someone joins, we give them information, they complete the required learning and at some point we declare them onboarded. The business, however, doesn't really care whether somebody is onboarded. It cares whether they can do the job.

Those two things are not remotely the same.

Completion is easy to measure. Competence is harder.

There are perfectly sensible reasons businesses measure onboarding through completion. It is visible, binary and wonderfully easy to put on a dashboard. Eight modules assigned, eight completed. Compliance training done. Systems access granted. Meetings attended. Perhaps an assessment passed at the end.

None of those things is useless, most are essential; but this process is incomplete if the aim is for fully ramped staff that produce more than they cost. The onboarding processes in place tell us considerably more about what the organisation has provided than about what the employee can now do.

That distinction becomes commercially important in any business hiring, especially if they are bringing in people at pace.

Imagine a sales organisation recruiting 20 people over the course of a year. If the business expects each of them to eventually carry a meaningful revenue target, every unnecessary week between joining and becoming independently productive has a cost. There is the obvious opportunity cost of the revenue they aren't yet generating, but there is also the less visible cost of the people around them. Managers answer questions, experienced colleagues join calls, work gets checked, mistakes get corrected and opportunities that matter are quietly routed towards safer pairs of hands.

This isn't an argument for rushing people through onboarding. Quite the opposite. Poorly prepared people being thrown into live work quickly is not an efficiency gain; it is simply moving the cost somewhere less visible.

The question is whether the journey has been designed around time-to-completion or time-to-competence.

There is a useful body of research here because academics have generally studied onboarding under the broader heading of organisational socialisation: the process through which somebody learns enough about their role, work and organisation to move from being an outsider to functioning effectively within it.

Bauer and colleagues' recent meta-analysis is particularly useful. Published in the Journal of Management in 2025, it reviewed 256 studies of newcomer socialisation and found that successful adjustment is not one thing. Social acceptance, role clarity and task mastery all matter, although in different ways. Most relevant to a performance conversation, task mastery showed a strong relationship with subsequent job performance.1

That seems obvious when written down. Yet a surprising amount of onboarding is still designed primarily around exposure to information rather than mastery of work.

We tell people about the company, the product or service offering, about processes and systems. Then, having successfully exposed them to everything, we become disappointed when they can't just perform.

Knowing about the work and being able to do the work are different states.

The problem usually becomes somebody else's workload

This is where onboarding becomes more than an HR or L&D issue.

I worked with one sales-led organisation where the official onboarding programme lasted several weeks, but the unofficial onboarding programme lasted months. The official version contained all the things the new starter was supposed to learn. The unofficial version consisted of experienced managers and high performers repeatedly helping them work out how to apply it.

Nobody had designed the second programme. It simply emerged.

Managers would listen to calls, rewrite emails, join meetings that technically shouldn't have required them, answer questions on Teams and occasionally take over an opportunity because the commercial risk of letting somebody learn through the mistake was simply too high. None of this appeared in the onboarding completion data. From the learning system's perspective, the employee had finished.

From the manager's perspective, they were still onboarding them every day.

One of the most useful onboarding measures is not something we measure about the new starter at all. It is the amount of support they still require from everybody else.

If somebody has completed every module but still needs their manager in every important client conversation, they aren't fully competent yet. If they know the process but can't recognise when the standard process doesn't apply, there is still development required. If they can pass the product assessment but can't explain the proposition convincingly to a sceptical prospect, the assessment has measured the wrong thing.

A 2023 systematic review by Frögéli, Jenner and Gustavsson is useful here, although it also gives us a reason to be cautious about grand claims concerning onboarding. The authors found that the evidence base for specific formal onboarding practices was weaker than we might expect, with only five studies meeting their criteria and the overall certainty of evidence rated low. The approach with the strongest support in that review was structured, supported on-the-job training.2

I think that's an important finding, partly because it challenges the temptation to solve onboarding by continually adding more content.

People need information, obviously. But they also need opportunities to perform representative pieces of the job, receive useful feedback and try again.

For a salesperson, that might mean conducting discovery rather than watching a video about discovery. For a recruiter, taking a properly qualified job rather than completing a module on qualification. For a new manager, handling a difficult performance conversation rather than demonstrating that they can remember the company's performance-management process.

The closer onboarding gets to the actual work, the more interesting our measures become.

Start at the other end

One of the simplest changes I make when looking at onboarding is to start at the end. Rather than beginning with “What should we put in the programme?”, begin with a much more demanding question:

What would this person need to be able to do for us to describe them as independently competent?

Not exceptional. Not finished developing. Not your top performer. Competent.

For a salesperson, perhaps they can independently run a discovery conversation, qualify an opportunity, navigate the CRM correctly, articulate the proposition, handle the common objections and know when they need help. For a new manager, perhaps they can set clear expectations, run a useful one-to-one, give difficult feedback and coach somebody through a performance problem without immediately taking the work back from them.

Once those outcomes are clear, the onboarding design becomes much easier to challenge. What does somebody need to know to perform them? What do they need to see demonstrated? What do they need to practise? Where do they need feedback? Which experiences can safely be simulated and which require real work? What support should remain available afterwards?

This also makes the role of managers considerably clearer.

The 2025 Bauer et al. meta-analysis found mentoring and support to be an important antecedent across newcomer outcomes, with relationships to performance and well-being as well as attitudes towards the organisation.1 Onboarding therefore cannot sensibly be something L&D or HR “delivers” before handing the finished employee to the manager.

The manager is part of the onboarding system.

That matters particularly in growing businesses because manager behaviour is often where beautifully designed onboarding falls apart. One new starter gets a manager who creates opportunities to practise, gives specific feedback and gradually increases autonomy. Another gets a manager who is busy, cancels their one-to-ones and answers every difficult question by taking over.

Both employees completed exactly the same programme. Their route to competence can look completely different.

Measure the distance to useful

If I were reviewing onboarding with a leadership team, I wouldn't begin by asking to see the content. I'd ask when a new hire becomes useful.

Then I'd make that more specific. When can they perform the core elements of the role without disproportionate supervision? Where do people consistently get stuck? What mistakes keep appearing after the relevant training has supposedly been completed? At what point does manager involvement begin to reduce? How much variation is there between managers, teams and cohorts? And, perhaps most importantly, what distinguishes the people who become competent unusually quickly?

Those questions tend to lead somewhere more interesting than another discussion about whether the induction deck needs updating.

They may expose missing knowledge. They may reveal that the learning isn't close enough to the work. They may show that managers aren't equipped to coach. They may identify a process that is unnecessarily difficult for newcomers to navigate. They may suggest that people need better access to information at the moment they use it, which is where an LMS, LXP, AI-enabled knowledge system or other performance-support technology might genuinely help.

But again, the solution comes after the diagnosis.

This is also where the economics become useful. You don't need a perfect financial model. Take a role you recruit regularly and establish roughly how long it currently takes before a new starter can perform its important activities independently and consistently. Then look at what happens during that period: the productivity you're waiting for, the manager time being consumed, the experienced colleagues providing support and the mistakes or missed opportunities associated with learning the job.

Now ask what would happen if you shortened that journey by two weeks. Or four.

In a business hiring three people a year, perhaps not much. In a business hiring 30, it starts to become a rather different conversation.

That's the point at which onboarding stops being something nice that happens to new employees and becomes what it should have been all along: a performance system designed to shorten the distance between hiring somebody and getting the value you hired them to create.

Completion still matters. Compliance still matters. People should understand the organisation they're joining and feel welcomed into it. In fact, the evidence suggests that social acceptance is a particularly important part of successful newcomer adjustment, so reducing onboarding to a cold productivity exercise would be just as misguided as reducing it to a content library.1

But if you're running a growing business, there is one question I would add to whatever onboarding dashboard you currently use. Not “Have they finished?”

What can they now do that they couldn't do when they arrived, and how long did it take us to get them there?

Once you start measuring that, you tend to have a much better conversation about onboarding. And usually, a rather more commercial one.

RB