This post in 30 seconds.
- The problem: a technician arrives without the right parts because the job was assigned from a sentence. The fault was described, the equipment was never identified, and the skill was a guess.
- The number: Aberdeen's field service research puts incorrect or missing parts behind 51% of second visits and the wrong skill behind another 25%. The fix respondents ranked first was better triage at dispatch, not more parts in the van.
- The model: collect the type plate, the serial, the fault code and a short guided video from the person already standing next to the equipment, then assign the technician and pick the parts from what came back.
- Who this is for: dispatch and planning leads, heads of field operations and service managers at equipment manufacturers, dealers, installers, utilities and maintenance providers.
The technician rings in from site at 10:40. He's standing in front of the unit, the customer is standing behind him, and the part in the van is the wrong revision. Somebody now has to explain a second visit to a customer who already took the morning off.
Ask how it happened and the answer is almost never "the warehouse was out of stock". The part was in stock, it just wasn't the part this machine needed, because when the job was assigned nobody knew which machine it was. The ticket said "not heating", a name, and a postcode.
If you run dispatch or planning for a field service operation, you've paid for that visit more than once this month. This post is about the three ways it happens, what to collect before you assign, and what changes on the planning board when you do.
We build Venta Capture, a product of VentaVid, so the person already on site can show your planner the plate, the serial and the fault before a van moves. Start for free and send a test flow to your own phone before you read on.
In this post:
The visit was lost before the van left
Most field service teams treat the wrong-parts visit as a stock problem. Bigger van stock, better forecasting, a parts run at lunchtime. Those help, and none of them touch the visit where the part was on the shelf and the wrong one went out.
Industrility describes the intake loop in one sentence: a customer describes a problem in plain language, a coordinator logs it in shorthand, and "unit not working" becomes the entire description a technician has to work from. The parts counter then picks from that description. So does whoever assigns the job.
The wrong-parts visit is an information problem first and an inventory problem second, and the information goes missing at intake. The numbers behind that claim come from one primary source that every vendor blog quotes and few have read.
Aberdeen Group surveyed 156 service organisations for its report Fixing First-Time Fix. Asked why a second or third visit happened, respondents answered:
- 51%: part unavailability, which Aberdeen defines as incorrect or no parts on site.
- 25%: the technician lacked the necessary experience for the job in front of them.
- 13%: insufficient time to complete the task.
Then Aberdeen asked what would fix it. The answer that came first, from 58% of respondents, was better diagnosis or triage at the dispatch or initial-call level, with improved access to parts in the van second at 33%. The people living with the problem ranked knowing more before dispatch above carrying more.
The same report has a finding that rarely gets quoted. Organisations that route all incoming calls through a triage step report an 86% first-time fix rate, against 62% for those with no triage at all. Aquant's 2025 benchmark across nearly 160 organisations shows the same spread from the other side: avoidable dispatch runs 3% at the top performers and 24% in the bottom fifth.
First-time fix as a metric deserves its own post, and it has one in this series. Here the question is narrower. What did the planner not know when the job went out, and how do you get it before the next one does?
The three ways a technician ends up on site with the wrong parts
Pull twenty repeat visits from last quarter and read the original ticket for each. Almost every one falls into one of three patterns, and the pattern tells you what was missing.
Cause 1: the fault was described, not shown
The customer says "it's making a noise" or "the screen is weird" or "it was fine yesterday". They are describing a symptom from memory, in a vocabulary nobody taught them, to a coordinator who needs a different set of facts entirely.
A heat pump logged as "not heating" turns out to be a tripped condensate float switch. A pallet truck logged as "won't lift" is low on hydraulic fluid. A coffee machine "leaking" is a split hose that the technician could have brought if anyone had seen where the water was coming from.
The technician packs for the sentence. When the sentence and the fault disagree, the part stays in the van and the van comes back.
Cause 2: the equipment was never identified before dispatch
This is the cause most advice skips, and it's the one that decides the parts pick. A model number identifies a family. A serial number identifies the variant, and the variant decides the part.
PRS Parts, which sells appliance components for a living, puts it plainly: manufacturers produce dozens of variations for each model line, and some parts look identical but function differently across models. Part numbers change with redesigns.
The same washing machine model can carry two drain pump revisions depending on the production run. A boiler's control board changes at a serial break. A forklift comes with three mast options under one model code.
Aberdeen's report carries a quote from the network manager of a European auto repair organisation that every planner should have on the wall: "In our industry product identification is the key to reduce repeat visits." Identification comes before diagnosis. If the serial number isn't on the ticket, the parts counter is guessing at spare parts identification from a family name.
Cause 3: the skill needed was guessed
A quarter of Aberdeen's second visits came down to a technician without the right experience. The usual response is skills-based routing in the scheduling tool, and it's the right response as far as it goes. The catch is that routing by skill needs to know which skill.
"Not heating" can be a refrigerant job, an electrical job or a controls job. Send an F-gas technician to a control board fault and you've paid for a competent person to confirm he can't fix it.
Send the mechanical technician to a hydraulic leak that turns out to be a sensor and the same thing happens in reverse. The technician levels in your matrix are only as useful as the ticket that triggers them.
"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."
Andy Huber, Manager of Systems Strategy, Xerox Technical Services, quoted in Aberdeen Group's Fixing First-Time Fix
All three causes share a root. The person who could have settled the question, the customer or operator standing next to the machine, was never asked for anything more than a sentence.
What removes the guesswork: five things to collect before you assign
The good news is that the evidence set is short. Five items, all of them capturable on a phone by someone with no training, in under three minutes.
Each row does one job. The plate gets you the family, the serial gets you the variant, the code gets you the subsystem, the video checks the story, and the surroundings photo prevents the third failure in Aberdeen's list, the visit that ran out of time.
"Send us a photo of the problem" doesn't get you this set; a guided request does, because the customer doesn't know what a serial break is and shouldn't have to. The difference between a photo in an inbox and a usable case is who decided what to capture. When your team builds the steps, the order and the instructions into the request, the right information comes back the first time.
That's what Venta Capture is built for. The planner sends a secure, personal capture link by SMS, WhatsApp or email, or drops a QR code into the booking confirmation. The customer opens it in the phone's browser, no app, no account, and walks through the steps your team defined: a scan field for the serial with an example photo of where the sticker sits and a typed fallback, photo slots for the plate and the display with a one-line instruction each, a yes/no question that steers the next step, and a video step with an overlay that tells them to walk around the unit and describe out loud what they see.
What lands in your inbox is a structured case, not a folder of images: the media, the answers to each question, a transcript of what the customer said while filming, server timestamps, and your own work order number in the reference field. One click copies it for the CRM or the FSM. If the plate is blurred, a retake request goes out in one click and the retake asks the customer why.
The guided capture flow does the expert's asking, so the expert only has to look.
I've watched the same dispatcher run both versions. With "can you send a photo of the problem?" the reply was a picture of a closed cabinet. With a guided request the reply was the plate, the serial typed in, a photo of an error code, and a video in which the customer pointed at where the noise came from, and the parts counter changed the pick before the van left.
The reason to do this asynchronously rather than on a live video call is capacity. A scheduled call needs a calendar slot and a staffed session for every ticket.
A link goes out with the confirmation, the customer captures when it suits them, and the planner reviews the case when it lands. Nobody has to be on the other end of the camera. If you already run remote diagnostics by phone, this is the visual layer under it.
If you'd like to see it against your own job types, book a demo and bring the three tickets that came back twice last month.
Assignment by evidence: what changes in planning
Once the evidence arrives before assignment, the planning board runs in a different order. The dispatch decision moves from the moment the ticket lands to the moment the case lands, and everything downstream gets easier.
- Ticket in, capture request out. The confirmation message carries the link. Nothing else changes for the customer yet.
- Case back before assignment. Plate, serial, code, video, surroundings. The planner reads it in under a minute.
- Parts picked from the plate, not the sentence. The counter sees the model and serial and picks the revision that fits. Aberdeen found the needed part was in the van 58% of the time at Best-in-Class organisations and 45% elsewhere; the gap is mostly whether anyone knew which part to load.
- Skill matched to the code, not the adjective. A control board fault goes to the controls technician. Skills-based routing finally has an input worth routing on.
- Time allocated from the surroundings photo. A unit on a mezzanine with no lift gets the slot and the second pair of hands it needs.
- Van leaves with the part, the skill and the time. The three causes in Aberdeen's list, closed before departure.
The planning consequence is that assignment becomes a decision made on evidence, and a repeat visit becomes an exception you can explain rather than a cost you absorb. Aberdeen puts the average at 1.6 extra dispatches for every job that fails on the first visit, and ServiceMax's field service survey, quoted by Salesforce, found 76% of customers had voiced dissatisfaction with a technician failing to resolve on the first visit. Every repeat you prevent removes both.
Some cases never need the van at all. A photo of a tripped breaker or a video of a blocked filter settles from the desk, and Venta Capture users report 23% of customer problems solved without dispatching a technician. That's a bonus, and it belongs to a different post on reducing truck rolls.
The point here is the visits that still happen: they leave prepared.
Plan the fallback before you start. Some customers won't capture, some links won't get opened, and a genuine emergency shouldn't wait for a video. When no case comes back inside your window, dispatch exactly as you do today, but tag the work order "dispatched blind" and count them.
After a month you'll have two populations of visits, and the repeat rate on each is the business case written for you.
How to run it for two weeks without touching your FSM
None of this needs a systems project. The link, the flow and the inbox sit next to your field service management tool, and the case gets pasted into it with one click.
- Pick two job types that keep coming back. Use last quarter's repeat visits. The ones where the second trip was "wrong part" or "wrong tech" are your candidates.
- Build one flow. Plate photo, serial scan, display or code photo, a 30-second video, a surroundings photo. Start from a standard flow and adjust the instructions to your equipment. If you want branching by answer or a route into your ticketing, that gets set up around your process on a call.
- Send the link with the booking confirmation. SMS or WhatsApp for consumers, email with a QR code for site contacts who'll be in front of the machine later.
- Route the case to the planner and the parts counter before assignment. Both look at it. The counter picks, the planner assigns.
- Count three things. Evidence-back rate (how many links came back with a usable case), picks changed after evidence (how often the part or the technician differed from what the ticket suggested), and repeat visits on those two job types against the previous quarter.
Good, after two weeks, looks like a parts counter who stops asking "which one have they got?" and a planner who can point at the plate when the assignment gets questioned. The same pre-visit assessment discipline that settles whether to send anyone also settles whom to send and with what.
If the technician then captures a proof of work on the way out, the job has a before and an after in one file, and the next planner who touches that asset can see what was fitted last time. That is how guided photo capture pays twice: once at intake, once at close.
Ready to try it on next week's queue? Start for free, build the five-step flow, and send it to your own phone first.
The identification step itself is worked out in equipment identification on video, and the metric it moves in how to improve first-time fix rate.
Frequently asked questions
Why do technicians arrive without the right parts?
Because the parts were picked from a description rather than from the equipment. Aberdeen's research found incorrect or missing parts behind 51% of second visits, and the fix its respondents ranked first was better triage at dispatch, not more stock in the van. If the plate and serial were never captured, the counter was guessing at the variant.
What information should a dispatcher collect before sending a technician?
Five things: a photo of the type plate, the serial number, the display or fault code, a short video of the fault, and a photo of the access and surroundings. Together they settle the parts family, the variant, the likely subsystem, whether the described fault is real, and how long the job needs. A guided capture request collects all five from the customer without a phone call.
What if the type plate is faded or the customer can't reach it?
Ask for a photo of whatever label survives, including any barcode, plus the control board or the installer's sticker; PICKHVAC's guide lists secondary labels, service stickers and the board part number as the usual fallbacks. Your service history for that site address often fills the gap. If none of that works, dispatch as today, tag the job "unidentified equipment" and count it.
Isn't a phone triage script enough?
It's a good start, and Aberdeen's 86% versus 62% first-time fix figure is for triage of any kind. A script still relies on the customer reading a serial aloud correctly and describing a noise in words. Adding a guided photo and video step removes the two places the phone version goes wrong.
Our FSM already does skills-based routing. What does this add?
An input worth routing on. Routing by skill needs the ticket to say which skill, and "not working" doesn't. A fault code and a 30-second video turn "not working" into "controls" or "refrigerant" before the routing rule runs.
What if the customer won't film or can't find the plate?
Some won't, and the flow should never block a genuine emergency. Dispatch those exactly as you do now, tag them "dispatched blind" and compare their repeat rate with the jobs that came with evidence. In practice a plate photo with an example picture of where the sticker sits gets a far better response than "send us the serial number".
Does collecting evidence before dispatch slow response times?
The link goes out with the booking confirmation and the customer captures in a few minutes, so for most jobs the case is back before the planner would have assigned anyway. The visit itself gets faster, because the technician arrives knowing the machine. The trade-off worth making is a short wait for evidence against a second trip.
Send the right technician with the right parts
Three causes, one root. The fault was described, the equipment wasn't identified, and the skill was guessed, all because nobody asked the person next to the machine for more than a sentence. Ask for the plate, the serial, the code and thirty seconds of video, and the assignment writes itself.
What it takes to start.
- Free plan, no credit card. Build the five-step flow for your most common repeat job and send it to your own phone first.
- Live in 10 minutes for a first standard flow. Stuck? Book a free setup call and we build it together.
- Nothing for the customer to install. The link opens in the phone's browser, in their language, with your branding on it.
Venta Capture is built by the VentaVid team, and a real person picks up the setup call.
Start for free and put the link on next week's "not working" tickets. If you'd rather see it on your own equipment first, book a demo and bring the three jobs that came back twice.