Why New Starters Lose Their First Week to Equipment

TL;DR

• The order is usually placed when onboarding starts, which in any market with a lead time over two weeks is already too late.
• Four gaps make up the total and only one of them is yours to control. Measure the acceptance-to-order gap, not the day-one pass rate.
• Three things break a first morning: an account created too late to enrol a second factor, a machine that arrives unconfigured, and nobody available to call.
• Zero-touch enrolment is the single biggest lever on whether a machine is usable when it is unboxed.
• Nobody tells the manager the machine arrived, which is why half of these failures are discovered by the new starter.
• When it will be late, say so and arrange something deliberate. Silence is the version that costs you the most.

Thirty-Five Minutes In

A company tracked onboarding satisfaction and had good scores. It also had a recurring comment it had not acted on, which was some version of the first week being slow to get going.

Somebody traced one starter's experience properly. The offer was accepted on the 3rd. Onboarding kicked off on the 24th, which is when the equipment request was raised, because that is where it sat in the checklist. The laptop was ordered on the 26th, in a market with an observed lead time of 19 days, for a start date on the 2nd of the following month.

The machine arrived on day 9. In the meantime the person had been given access to a shared document, watched some recorded sessions on a personal device, and had two coffees with their manager. Every individual step had worked and the one controllable number, the 23 days between acceptance and order, had never been looked at by anybody, because no report measured it and no person owned it.

Recommended Tools

Best tools for HR Operations

B BambooHR HR management system R Rippling HR & IT automation W Workday Enterprise HRIS

View all recommended tools

The fix was moving the order trigger from the onboarding kickoff to the acceptance date. It cost nothing, required one notification to be wired, and removed 23 days from the critical path.

The Four Gaps

The total wait is made of four intervals, and companies routinely try to fix the wrong one.

The first row is the one nobody counts, because it is not experienced as a delay by anybody. The person is not working yet, so no clock is visibly running, and the interval sits in a part of the process that belongs to recruiting rather than to onboarding.

That gap is usually larger than the supplier's lead time, which means most companies are trying to negotiate a faster delivery while sitting on three weeks of their own slack. Measuring it is one query against two dates.

The One Date That Fixes It

Order at acceptance.

Not at onboarding kickoff, not at the start of the pre-start process, not when the manager remembers. At the moment somebody accepts, which is the earliest date on which you know the person, the role and the market.

Three things have to be true for this to work.

Somebody has to be told, with three facts. The person, the market and the role. Nothing else. This is a small notification from a system holding sensitive data, so specify the payload rather than exporting the record.

The specification has to be known without a conversation. Which means the tier sits on the role definition rather than being decided by a manager after acceptance. If it needs a conversation, the order waits for that conversation and you have reintroduced the delay in a smaller form.

You have to accept ordering for people who do not start. Offers fall through. A company ordering at acceptance will occasionally hold a machine nobody needs, and that machine is the first item in a spare pool rather than a loss. Treating it as waste is exactly what pushes companies back to ordering late.

So the trade is a small number of unused machines a year against day-one readiness in every market. Because the unused machine is visible and the nine-day wait is not, most companies have implicitly chosen the worse side of that trade without anybody deciding it.

What Day One Actually Needs

Five things, and the machine is only one of them.

A machine that is configured. Not a machine that has arrived. The difference is zero-touch enrolment, and without it somebody spends their first morning on a setup call.

An account that exists, with a second factor already enrolled. This is the one that fails most and surprises people. A second factor usually needs enrolling from a device, and if the account was created the night before, nobody has done it. The person then cannot sign in to the thing they need in order to enrol the second factor, which is a loop somebody has to break manually.

Access to the specific things they need on day one. Not everything. The three or four systems they will actually open, granted and tested.

A named person to contact, who is available. Not a queue name. A person, with a direct message handle, who knows a new starter is beginning today.

Somewhere to start. A short list of what to do in the first few hours that does not depend on anything above going perfectly, so a hitch is an inconvenience rather than a vacuum.

The gap between a good first day and a bad one is rarely the hardware. It is usually the second factor and the named person.

When Day One Readiness Is Not the Right Target

When the start date moved and nobody updated anything

Start dates shift, and a machine ordered against the old date can arrive weeks early, which is fine, or an order placed against a date that moved earlier can be late through nobody's fault. Make the start date change a trigger that re-checks the equipment, because it currently triggers nothing.

When the market genuinely cannot support it

If the lead time in a market is 5 weeks and notice periods there are 2 weeks, no process change will produce a day-one machine. That is a spare-pool problem rather than a scheduling one, and pretending otherwise produces a promise that fails every time.

When the role does not need a machine on day one

Some roles genuinely begin with induction, shadowing and reading. If that is true, say so and plan for it, because an unplanned equipment gap and a planned induction week look identical from the outside and feel completely different.

The edge case that forces the full treatment

A cohort. Five or ten people starting on the same day turns every small scheduling weakness into a visible failure, and it is also the easiest case to plan because the date is known months out.

Five Questions People Ask First

"Why not just order earlier in the onboarding process?" Because the onboarding process starts too late. Moving the order to the first day of onboarding saves a few days; moving it to acceptance saves a few weeks. The second is the change worth making and it requires a notification rather than a process redesign.

"Can we not just ship it to arrive on day one?" You can, and arriving on day one is worse than arriving a week early. A machine that arrives the morning somebody starts cannot be checked, and if anything is wrong you discover it with the person sitting there. Aim for 3 to 5 days early, which gives you a window to fix a problem invisibly.

"What if we do not know the address?" Ask for it at acceptance, in the same message that confirms the start date. This is the natural moment, the person is engaged, and it is considerably easier than asking later. Also ask for a phone number, because couriers need one and a failed delivery costs days.

"Who should own this?" One named person, and the honest answer at most companies is that it should not be the hiring manager. The manager is the least well placed to chase a supplier and the most likely to assume somebody else is doing it. Whoever owns onboarding is usually the right owner, provided they receive the acceptance notification.

"Does zero-touch enrolment really matter?" More than anything else on the hardware side. The difference between a machine that configures itself when unboxed and one that needs a 90-minute call is the difference between a first morning and a first afternoon, and the constraint is that the device must be bought through a channel that supports it. That single requirement should drive supplier selection ahead of unit price.

The Second Factor Problem

The failure that accounts for more bad first mornings than hardware does, and the one that is almost never on a checklist.

The pattern is consistent. The account is created a day or two before the start date, which feels prudent. Nobody enrols a second factor, because that is something the person does. On the morning, the person tries to sign in, is prompted to enrol a second factor, and cannot complete it because the enrolment flow requires them to be signed in, or requires a code sent to an address they cannot yet reach, or requires an administrator to issue a temporary method.

Three fixes, and the first is nearly free.

Create the account at least 5 working days early and issue a temporary enrolment method. A time-limited code or a pre-registration link that the person uses before their first day, sent to their personal address along with the pre-start email.

Have the named contact available at the start time with the ability to issue a reset. Not a queue. The person who can unblock this in two minutes, online, at the moment the new starter is trying.

Test the whole flow on a spare account once a quarter. Sign in as a brand new user on a brand new machine and see what happens. Because nobody in the company ever does this except new starters, the flow is tested only by the people least able to diagnose it, and it changes whenever the identity configuration is updated.

And write down what the person should do if it fails, in the pre-start email, including the contact. A new starter who cannot sign in and has a name to message has an awkward ten minutes. One who cannot sign in and has nobody to ask has a bad day and tells people about it.

Nobody Tells the Manager

A small gap with a large effect: the machine arrives, and the only person who knows is the courier's tracking system.

The consequences are predictable. The manager does not know whether to plan a working first day or an induction day. Nobody checks the machine before the person opens it. And if it has not arrived, the manager finds out at the same moment the new starter does, which is the worst possible sequencing.

Notify the manager and the onboarding owner on delivery, automatically where the courier integration allows it and manually where it does not. One line: the machine for this person was delivered on this date.

Then have somebody check it. Powered on, enrolled, encrypted, signed in once as a test, and the keyboard layout is the one that was ordered. 10 minutes, 3 to 5 days before the start date, and it catches the problems that would otherwise be discovered live.

And escalate on the exception rather than reporting the rule. A notification that says nothing has been delivered and the start date is in 4 days is worth more than a weekly report nobody opens. Exception alerting is the whole value here, because the successful cases need no attention at all.

When It Will Be Late

Sometimes it will be, and the difference between a handled delay and an unhandled one is almost entirely communication.

Tell the person before they start, not on the day. In the pre-start email, with the expected date. Somebody who knows their machine arrives on day 4 can plan; somebody who finds out on day 1 spends the first morning in uncertainty.

Arrange something specific for the interim. A loaner from the spare pool if one exists. Access through a browser on a personal device if the work and your security position genuinely allow it, which is a decision to make deliberately rather than by default. Or an induction-weighted first few days, planned rather than improvised.

Do not let a personal device become the arrangement by accident. This is the quiet risk in every late delivery. Somebody who cannot work waits a day and then uses their own laptop, which puts company data somewhere you have no visibility of and no way to clear. If that is acceptable to you, decide it and set up the access accordingly. If it is not, say so and give them something else to do.

Tell the manager the revised date too. The manager is planning somebody's first week and is the last to be told.

So the honest position is that a late machine handled well is a minor inconvenience, and a late machine handled by silence produces both a bad first impression and an unmanaged data risk. The handling costs two messages.

How to Choose: Five Questions Before You Change Anything

What is your acceptance-to-order gap, in days, by market? The single number this whole article rests on. Two dates, one query, and it is almost always larger than the supplier lead time you have been trying to negotiate down.

Is the tier on the role definition before the offer goes out? If not, every order waits for a conversation and the trigger change will not deliver what it should.

Who receives the acceptance notification? If the answer is nobody, that is the gap. Three fields, one recipient.

Are your devices bought through a channel that supports automated enrolment? If not, every machine needs a manual setup call regardless of when it arrives, and the readiness problem is only half solved by fixing the timing.

When did anybody last sign in as a brand new user on a brand new machine? If the answer is never, that flow is being tested exclusively by new starters.

The Options in Practice

Your own ordering, through an enrolled supplier

The baseline and the right answer in markets you already serve. One supplier relationship, zero-touch enrolment, and full control of the specification.

Where it struggles: markets you cannot reach, and residential delivery where you have no local presence.

A regional spare pool

Not an alternative to ordering, a complement, and the only thing that works where the lead time exceeds the notice period. Funded from recovered machines, held where you hire most, and it converts an impossible market into an ordinary one.

Where it struggles: it has to exist in advance, and it only helps the markets you chose to stock.

A lifecycle platform

Best where several markets are involved and you want purchase, configuration and delivery as one step with a collection arrangement at the other end.

Workwize, Deel IT, Firstbase, GroWrk and allwhere all operate here and none publishes a price, so comparing them means several sales conversations before you have a figure. The capability to test for this use case is whether a machine arrives enrolled and ready rather than merely delivered.

Disclosure: RemoAsset is owned by the same people who publish HROpsLab. It is relevant to day-one readiness specifically because the purchase, the configuration and the shipment are a single step, which removes the delivery-to-usable gap rather than shortening it, and because the order captures the serial and the address at the only moment they are reliably available. It publishes no price and requires a demo, it is much weaker for hardware it did not supply, and it is not a certified disposal vendor.

The Comparison

The second row is the cheapest change on the page and the fourth is the only thing that works where the arithmetic is impossible. Most companies need both and have neither.

The Decision Table

Measuring the Gap You Control

The acceptance-to-order gap is the number this article rests on and almost nobody has it, so here is how to produce it in an afternoon.

Take every person who started in the last two quarters. Not the last year, because process changes make older data misleading, and not the last month, because the sample is too small to show the variation by market.

Pull two dates each. The offer acceptance date from the recruiting system, and the order date from purchase records or the asset register. Both exist; neither is usually in the same place.

Subtract, and group by market. You are looking for the distribution rather than the mean, because the mean will be dragged around by a handful of unusual cases and the thing you need to see is whether a particular market is consistently bad.

Three patterns and what each means.

A long gap everywhere, with low variance. The trigger is in the wrong place in the process, which is the easy fix and the common finding. One notification resolves it.

A short gap in most markets and a long one in two. Somebody owns the ordering in the good markets and nobody owns it in the others. This is a staffing or assignment question rather than a process one.

High variance everywhere. The ordering depends on an individual remembering, which means performance tracks how busy that person was that week. The fix is the same notification, plus an exception alert when a gap exceeds about 10 days.

Then put the supplier's observed delivery time next to it, by market, and the comparison usually settles the argument. Because a 23-day internal gap sitting next to a 19-day supplier lead time makes the priority obvious in a way that no amount of discussion about supplier performance does, this single table is worth more than any process document you could write about it.

And re-run it two quarters after the change, with the same method. A one-off measurement proves a problem existed; two prove the fix worked, and the second is the one that keeps the arrangement in place when the person who set it up moves on.

The First Thirty-Five Minutes

Worth defining what success looks like, because day-one readiness gets measured as whether a laptop was present and that is a weak test.

The useful standard is whether somebody can do a piece of real work in their first half hour. Not watch a video, not read a handbook. Open the thing they were hired to open and change something.

Six steps, and the whole sequence should take about 35 minutes including a coffee.

Open the lid and sign in. The machine configures itself, applies the settings, and the person signs in with the credential and the second factor they enrolled last week.

The applications are already there. Not a software catalogue to browse. The 6 or 8 things their role uses, installed, signed in where single sign-on allows it.

A message from the named contact is waiting. Sent that morning, before the person started, saying who they are and how to reach them. This costs one minute and changes the whole texture of the morning.

The 3 or 4 systems they need are accessible, and somebody tested them. Tested with their actual account rather than assumed from a group membership, because group-based access that has not propagated looks identical to access that was never granted.

A first task exists, small and real. Something with a visible result, chosen by the manager in advance. The difference between a good first day and a pleasant one is usually whether anything was produced.

And a fallback list that needs none of the above. Reading, recorded sessions, people to meet. Not as the plan, as the thing that absorbs a hitch without the morning becoming a vacuum. Every company has this list and most treat it as the first day rather than as the contingency, which is the distinction worth making explicit to whoever writes it.

So measure the 35 minutes rather than the delivery date. Because a machine can be present and the morning can still be lost to an access request, the delivery date is a necessary condition masquerading as the outcome, and companies that track only the hardware report success on days their starters remember badly.

Working Backwards From the Start Date

One practical artefact worth having, because the four gaps are easier to manage as a countdown than as a process.

At acceptance. Order the machine. Ask for the delivery address and a phone number in the same message that confirms the start date. Record the tier from the role definition.

Twelve working days out. Confirm the order is on track and the expected delivery date still clears the start date with margin. This is the last moment an alternative can be arranged without drama.

Ten working days out. Create the account. Issue the temporary second-factor enrolment method to the person's personal address.

Five working days out. Send the pre-start email: the machine by name, when it arrives, what they do on arrival, who to contact, and the second-factor enrolment link if they have not used it.

Three working days out. Machine should be delivered. Check it: powered on, enrolled, encrypted, one test sign-in, keyboard layout as ordered. Notify the manager.

One working day out. Confirm access to the 3 or 4 day-one systems using the person's own account. Manager confirms the first task exists.

Day one, 15 minutes before the start time. The named contact sends their message.

Six checkpoints, each a few minutes, and the only one that cannot be moved is the first. Because every checkpoint after acceptance depends on the order having been placed, the countdown either works or it does not from the very first step, which is why the acceptance trigger carries more weight than the other five put together.

What Getting This Wrong Costs

The visible cost is a slow first week, and most companies know they have it because somebody writes it in a survey.

The second cost is the impression, and it is not recoverable in the way a later improvement might be. Somebody's first week is the only one they experience as a new person, and a company that could not arrange a laptop has said something about itself that no amount of later competence fully undoes. Because the comment arrives as a mild remark in onboarding feedback rather than as a complaint, it reliably gets logged and not acted on, which is how the same failure persists through several years of good satisfaction scores.

The third is the unmanaged data risk, and it is the one that should worry somebody. A person who cannot work waits a day and then uses their own machine, because they are new and want to make a good impression. Nobody authorised it, nobody knows, and there is now company data on a device you cannot see, cannot clear and have no record of. That is a direct consequence of a late order and it appears in no equipment report anywhere.

The reframing question: of the people who started in the last two quarters, how many could do real work in their first 35 minutes, and for the ones who could not, what were they doing instead?

When You're Ready to Fix It

The honest position is that this is one measurement and one notification, and it stays broken because the expensive interval is invisible to everybody involved.

The trigger worth acting on is a single traced starter. Pick one person who started recently, find their acceptance date and their order date, and subtract. That number is the business case and it takes five minutes to produce.

The sequence costs almost nothing. Measure the acceptance-to-order gap by market and publish it. Put the tier on the role definition so an order needs no conversation. Name whoever receives the acceptance notification and specify its three fields. Create accounts 5 working days early with a temporary enrolment method sent to a personal address. Then notify the manager on delivery and check the machine 3 days out.

And once a quarter, sign in as a brand new user on a brand new machine yourself. It takes twenty minutes, it is the only test of the flow that is not being run by somebody on their first day, and it is the one that finds the problems the checklist cannot see.

One thing worth separating from the mechanics. There is a tendency to treat this as an IT matter, because the visible objects are a laptop and an account, and that framing is why it stays broken. Every expensive interval in this article sits between two functions: acceptance to order sits between recruiting and procurement, account creation sits between HR and IT, and the day-one check sits between IT and the manager. None of the four gaps belongs wholly to anybody, which is exactly why each one has quietly grown to whatever length nobody complains about.

So the structural fix is not a better checklist. It is one named owner for the whole span from acceptance to the first thirty-five minutes, with the authority to ask three other teams for a date. That person does not need to do any of the work. They need to be the one who notices when a gap is 23 days, which is a role nobody currently holds at most companies and the only reason the number has never been looked at.

And if you take one thing from this: find one recent starter, subtract their acceptance date from their order date, and look at the answer. Everything else here follows from whether that number surprises you.

Frequently Asked Questions

HROpsLab takes no vendor money and publishes no paid placements, which is why the owner's own product is named above with its three limitations rather than its three best features.