This post in 30 seconds.
- The problem: most repeat visits are decided before the van leaves, off a ticket that says "unit not working" and a part list guessed from it.
- The number: a failed first visit adds two more visits and 14 days to the job, according to Aquant's 2025 benchmark of nearly 160 service organisations.
- The model: agree what counts as a fix, treat the four causes of a repeat as four missing pieces of information, and collect them from the person already on site before you assign anyone.
- Who this is for: service managers, heads of field operations, dispatch leads and technical-support managers at equipment makers, dealers, installers, utilities and maintenance providers.
Look at any Tuesday board and there it is: the same job, second time this week. The note from the first visit reads "part on order", or "needs an electrician", or "no access, customer at work". Three different notes, one cost, and a first-time fix number that everyone reports and nobody quite trusts.
If you run a service desk, a dispatch team or a field operation, you already know the shape of the problem. Every guide on how to improve first-time fix rate tells you to buy scheduling software, stock the vans better and train the technicians. All fair. None of it helps when the job was assigned off one sentence a customer typed into a form, and the technician is diagnosing from scratch in the driveway.
This post takes the other route: get the definition straight, name the four reasons a technician goes back, show that all four are information you could have had before assignment, and lay out what to collect from the customer and from the technician so the second visit never gets raised.
If your dispatchers are sending people out blind and your repeat rate proves it, we built Venta Capture, a product of VentaVid, so the person standing next to the fault can show it to you, guided step by step, before you commit a technician. Start for free and send a test link to your own phone this afternoon.
In this post:
Agree what counts as a fix before you try to move the number
IBM's definition is the cleanest one in circulation: first-time fix rate is the percentage of jobs completed on the first visit "without requiring follow-up visits, supplementary parts or external support", and the formula is jobs fixed first visit divided by total jobs, times 100 (IBM). Simple on paper. In practice, two service managers in the same company will count the same job differently.
So before you spend a quarter improving the number, spend an afternoon agreeing what it measures. Three decisions do most of the work.
The strict version gives you a lower number, once. It also gives you a number that means something. When a technician replaces a board, tests OK and gets called back nine days later for the same fault, the loose version calls that two successful jobs. Your customer calls it one job that took two visits, and the top field-service complaint in Aberdeen's research, cited by 57% of respondents, was exactly that: "technician did not resolve issue" (Aberdeen, 2013).
A note on that source, because it turns up on every page that ranks for this keyword. Aberdeen surveyed 156 field-service organisations in early 2013, so the numbers are old. They are also the only cause-level data set most of the industry has, and the newer benchmarks (Aquant's, below) point the same way. I quote it with the date attached.
Aberdeen's average first-time fix was 75%, with the best performers at 89%. The other 25% of jobs needed at least one more visit, and those jobs averaged 1.6 extra dispatches each. Count your repeat visits that way and the callback rate stops being a footnote.
The four reasons a technician goes back
Ask the technicians why a job took two visits and you get four answers. Aberdeen's respondents put them in this order: the wrong part or no part (51%), a technician without the necessary experience (25%), and not enough time on site (13%). Sit with the tickets rather than the survey and a fourth one shows up constantly: nobody could get to the equipment.
- Wrong diagnosis. The ticket said "making a noise". The desk assumed a bearing. It was a loose panel and a mounting bolt, which the customer could have shown in ten seconds. The technician spent forty minutes finding it, and the fix needed a part that wasn't on the van.
- Wrong part. The model was right, the revision was wrong. Two boards look the same until you read the serial plate, and nobody read the serial plate before dispatch because nobody asked for it.
- Wrong skill. The job needed a refrigerant-certified engineer, or an electrician, or someone who has done that generation of controller. The scheduler picked whoever was nearest.
- No access, or not ready. Locked plant room. Machine still in production and not released. The customer's own contractor hadn't finished the cabling for the install. The van went anyway.
Every one of those four is a fact that existed before the job was assigned, and none of them needed a technician to discover. The panel was loose on Monday, the serial plate said what it said, and the plant room was locked at 9 a.m. either way. Parts sits at 51% in the survey because a wrong diagnosis and a misread identifier both get filed under "didn't have the part" when the technician writes it up.
That is why the van stock conversation is the wrong first conversation. Better parts availability helps once you know which part. Until then you are stocking the van for a guess.
What a repeat costs has been measured more recently. Aquant's 2025 benchmark, built on 600,000 technician service records, found that a failed first visit adds two further visits on average and stretches resolution by 14 days; top performers ran a first-time fix rate of 86% against 53% for the bottom group (Aquant, 2025). Aberdeen priced a single truck roll at $200 to $300 in operating cost. We covered how to cut the rolls that should never have happened separately; this post is about the ones that have to happen, and making them count once.
Why the fix starts at the desk, not in the van
The same Aberdeen report asked respondents what they planned to do about it. Every page that quotes the causes skips this chart, so here it is: 58% put "better diagnosis or triage at the dispatch/initial call level" first. Parts access came second at 33%, smarter scheduling at 32%, training at 25%. The practitioners knew where the leak was in 2013.
They had the data to back it. Organisations that routed every incoming call through a triage step reported 86% first-time fix. Those that triaged fewer than half of their calls reported 71%. No triage at all: 62%.
"There are two primary ways to improve first-time fix: 1. Accurately diagnose the problem and solution before going on site. 2. Bring enough parts to solve the most problems on one visit."
by Andy Huber, Manager of Systems Strategy, Xerox Technical Services, as quoted in Aberdeen's Fixing First-Time Fix (2013)
So triage works. The catch is what triage usually consists of: a phone call, a form, and a dispatcher with a good ear. The customer describes a noise. The dispatcher pictures a different noise. The best remote triage person in the building is still working from a description, and the person who can see the equipment doesn't know which of the forty things in front of them matters.
That gap is what guided capture closes. Instead of "send us some photos", the organisation builds the steps: scan the plate here, film the display now, walk around the unit and say out loud what you hear, answer these three questions. The customer, operator or site contact does it on their own phone in the mobile browser, no app and no account, whenever suits them, and the desk reviews a structured case when it lands. Send now, capture later, review when ready. It's asynchronous, which matters more than "remote": nobody has to book a video call to look at a control panel.
We have written up pre-visit assessment and remote diagnostics as workflows in their own right. Here, the point is narrower. If the four causes are four missing facts, triage is the step where you collect them, and a guided capture is how you collect them without a technician.
Want to see what a triage case looks like when it arrives with the plate, the display and the sound already in it? Book a demo and bring one of your own repeat tickets.
What to collect before you assign the job
The capture list is short on purpose. Every item on it either changes who you send, what they bring, or whether you send anyone at all. Anything that doesn't do one of those three is noise, and noise is why customers abandon long forms.
Two things about that table. The serial number row is first because it is the cheapest and does the most: the network manager of a European auto-repair organisation told Aberdeen that "product identification is the key to reduce repeat visits", and anyone who has ordered the wrong revision of a board will agree. Spare parts identification starts with a legible plate, and a scan step with an example photo of where the sticker sits gets you a legible plate.
The second thing is that the list only works if the customer follows it in order, with the right framing on each step. That is the difference between a photo request and a guided workflow. In Venta Capture's guided video capture each step carries its own plain-language instruction and an example image, the video step shows an overlay while the customer films ("walk around the machine and describe out loud what you see"), a yes/no answer steers which step comes next, and spoken explanations are transcribed so the desk can search "grinding" across a week of cases. What comes back is one case with the video, the photos, the answers and the timestamps tied to your own reference number, not six attachments in a WhatsApp thread.
Before you build anything, pull a month of your own second visits and code each one against those five rows. Expect a dull pattern, which is the useful kind: most of them trace back to one row, and in most of those the customer was standing within arm's reach of the answer on the day they called.
With Venta Capture, 23% of customer problems are solved without dispatching a technician at all, because the desk can see that the breaker tripped, the filter is blocked or the reset works. The jobs that still need a van go out with a diagnosis, a part number and a named skill instead of a sentence. The guided photo capture post goes deeper on how to design the steps; the minimum evidence set glossary entry covers how little you can ask for and still get a decision.
"Our customers won't do a video." Some won't. Most will, when it takes a few minutes, needs nothing installed, and the alternative is a two-week wait for a visit that may not fix it. The comparison isn't video versus nothing. It's a guided capture versus the phone call that produced "unit not working".
The technician's side: capture the condition found and the work done
First-time fix is also decided at close-out, and this is the half the ranking guides leave out. A work order that ends in "tested OK" with nothing behind it is how a callback turns into an argument about whether the first fix was a fix.
Two captures on site, both short, both guided the same way the customer's was:
- Condition found on arrival. Does the equipment match what the customer sent? Same fault, same display, same serial? Thirty seconds of video, and a yes/no that flags the mismatch to the desk before the technician starts.
- Work done before leaving. The part fitted with its serial, the reading after the fix, the test the technician ran, the customer's own confirmation on the same phone if you want it. Staff can start a capture themselves from the dashboard without sending a link, so this is a habit, not an admin task.
A fix you can't show is a fix you can't defend, and it is also a fix the next technician can't build on. Three things change once the technician's capture is routine. The first-time fix number becomes auditable: "fixed" means there is a sealed record of what was fixed, timestamped on receipt. The callback dispute has evidence on both sides, because the proof of work sits next to the customer's original capture in one file. And a repeat, when it happens, opens as a case linked to the original, so the second technician starts from the first technician's video instead of from zero.
We wrote a full guide to proof of work in field service, covering contractors and sign-off. For the metric, the point is simpler: technician efficiency improves when the record of the last visit is something you can watch, not something you have to ring the technician about.
Carry the evidence through to the van
Collecting the right information at the desk and then leaving it at the desk is the most common way this goes wrong. Salesforce, citing a ServiceMax field-service survey, put it at 38% of service agents saying they can't access all the information they need on the job, and 76% of customers voicing dissatisfaction after a failed first visit (Salesforce). The two numbers are related.
What the technician should have on the phone before knocking:
- The customer's capture. The video with sound, the display, the plate, the access route. Watched in the van, not summarised by a dispatcher.
- The answers. The four or five steering questions, so "already reset, still faulting" is known before the first conversation on site.
- The dispatcher's diagnosis and part pick. With the reasoning, so the technician can disagree early rather than late.
- The previous case, if it's a repeat. Linked to this one, with the last technician's condition-found and work-done captures.
Getting it there is a workflow question, and the answer depends on what your technicians already carry. Aberdeen listed the same four items in 2013 as the point-of-service information that moved the number most: parts visibility, service history, resolution steps and step-by-step instructions. A field service management system holds two of those; the customer's evidence usually never makes it in.
In Venta Capture every case has a view link you can hand to a technician who has no login, a "copy for CRM" button that pastes the case into the job in whatever system holds it, and status-driven workflow automation that forwards a case to your ticketing or FSM tool through webhooks and the API. Routing rules and conditional forwarding are set up around your process by us, not something to knock together on a Friday. Notifications go per route, immediately or batched, so the planner sees the case when it lands and the technician sees it when they're assigned.
If you are also weighing up a dedicated tool for this, our guide to remote inspection software lays out what to look for. If you would rather see it than read about it, start for free: the free plan needs no credit card, the first flow takes about ten minutes to build, and requesting a video by link from your own phone is the quickest way to see what a technician would get.
How to tell whether it worked
Track first-time fix alongside four other numbers, or you'll improve the headline figure by redefining it rather than by fixing anything.
- Repeat visits per 100 jobs, by cause code. Use the four causes as the code list: wrong diagnosis, wrong part, wrong skill, no access or not ready. Every second visit gets one code, entered by the technician at close-out.
- Avoidable dispatches. Jobs closed without a visit, as a share of all jobs. Aquant's top performers run a 3% avoidable dispatch rate against 24% for the bottom fifth.
- Share of jobs assigned with a capture attached. If this number is low, the rest won't move. It is the leading indicator.
- Callbacks inside your window. Whatever window you agreed in the first section, count against it every week.
Aberdeen's satisfaction table is the reason the finance director will care. Organisations above 80% first-time fix reported 87% customer satisfaction; those under 50% reported 46%. Retention followed the same curve, 88% against 60%. Nobody hits those bands by stocking the van harder. They hit them by knowing what the job is before anyone drives to it.
Two of the four causes get their own posts: why technicians arrive without the right parts and equipment identification before dispatch.
Frequently asked questions
What is a good first-time fix rate?
Aberdeen's 2013 average was 75% with best-in-class at 89%; Aquant's 2025 benchmark puts top performers at 86% and the bottom group at 53%. Above 80% is the band where satisfaction and retention jump. Compare against your own strict definition, not against a vendor's loose one.
How do you calculate first-time fix rate?
Jobs resolved on the first visit divided by total jobs completed, times 100. The arithmetic is easy; the decision is what "resolved" means and how long a callback window you count. Agree both in writing before you compare quarters.
Does a return visit for a part count as a failed first-time fix?
Under a strict definition, yes: the job needed a supplementary part and a second trip, which is the largest cause of repeats in the data (51% in Aberdeen's survey). Counting it as a success hides exactly the problem you are trying to fix. Code it as "wrong part" and look at how many trace back to an unread serial plate.
Will customers really record a video before a visit?
Most will when it takes a few minutes, runs in the phone's browser with no app or account, and every step tells them exactly what to film. Guided capture removes the "what do you want me to show you" problem that makes people give up on open photo requests. The alternative they are comparing it to is a wait for a visit that may not fix it.
Can remote capture replace a diagnosis?
No, and be wary of anything that claims it can. Guided capture puts the plate, the display, the sound and the surroundings in front of the person who diagnoses, so they are working from evidence instead of a description. The judgement stays with the qualified person; it just happens before the van moves.
What is the fastest way to raise first-time fix rate?
Add the equipment identifier and a filmed display to every ticket before assignment. Those two items alone take out most wrong-part and wrong-diagnosis repeats, and they cost the customer under a minute. Then add cause codes to every second visit so you can see what's left.
Talk to us
If the same job keeps landing on the board twice, the fastest way to see what changes is to send a capture link to your own phone and open the case that comes back. Then send it to one real customer with a real fault and compare it to the ticket you would have dispatched from.
What it takes to start.
- A free plan, no credit card. Build your first pre-visit flow in about ten minutes.
- No app for your customers. They open a secure, personal capture link in the browser and film.
- Stuck? Book a free setup call and we'll build your first flow around one of your repeat jobs.
The VentaVid team builds Venta Capture for service teams that need to see before they send someone.
Book a demo and bring a repeat ticket. We'll show you what the technician would have known.