How Many Accounts One Technical Service Engineer Can Cover
Summary
No standard ratio exists. An engineer can cover as many accounts as their available account hours divided by the visit, inquiry and test hours each account needs per year, weighted for complexity. The result holds only for the account mix you measured: an engineer with mostly repeat-order accounts carries several times more than one handling open failures or trials.
The workload model attributed to W. J. Talley divides the total workload of covering a market by the workload one person can handle to get the headcount needed (INFLIBNET Centre, accessed 2026).
In that model's textbook example, 9,100 hours of yearly call work divided by 1,056 selling hours per person gives 8.617, rounded up to 9 people (INFLIBNET Centre, accessed 2026).
The same example plans 1,920 work hours per person per year, 40 hours for 48 weeks, of which 55% goes to selling tasks (INFLIBNET Centre, accessed 2026).
GitLab's handbook page for assigned customer support engineers, modified October 9, 2026, gives each engineer a defined book of accounts bounded by capacity and a guide of around three engagements per account per quarter, with no fixed account count (GitLab, 2026).
This method sizes the account load of technical service engineers at a chemical or lubricant supplier who handle visits, inquiries and test work for named accounts. It does not apply to pooled helpdesk queues with no account ownership, to sales headcount set by revenue targets, or to field crews paid per job.
How do you calculate how many accounts one technical service engineer can cover?
The number of accounts one technical service engineer can cover is their available account hours per year divided by the weighted hours each account needs per year, summed over the account mix. The logic comes from the workload model for sales territories: management works out the total workload of covering the market and divides it by the workload one person can handle (INFLIBNET Centre, accessed 2026), an approach Walter J. Talley Jr. described in a case study of territory design for a national company (Talley, 1961).
Written for technical service, the formula is: accounts covered = available account hours / average weighted hours per account. Available account hours are what remains of the working year once leave, training, internal meetings, documentation and travel come out. The textbook example plans 1,920 hours a year and keeps 55% of them, 1,056 hours, for selling work (INFLIBNET Centre, accessed 2026). A technical service team should measure its own share from time records instead of copying that split.
Weighted hours per account are the hours in the service records multiplied by a complexity weight. ISO 9001:2026, the sixth edition, keeps people as a resource clause, 7.1.2 People (ISO, 2026), but the standard gives no figure: the staffing number is a decision each organization makes.
How do you count visit, inquiry and test workload per account from service records?
Pull twelve months of closed work for each account from the service records and convert every record into hours by work type. Three record types cover the work. Visits: count site visits and multiply by the average time per visit, report writing included, but keep travel separate so you can handle it by territory. Inquiries: count the technical questions answered by phone, email or chat, and use logged handling time where it exists; where it does not, time a sample of two or three weeks and apply the average. Tests: count the samples, used-oil or fluid analyses, compatibility checks and trials the engineer arranged or interpreted, and use the engineer's own time per test, not the laboratory's run time. Added together, the three give recorded hours per account per year.
Clean the records with two rules before you add anything up. First, merge duplicate records for the same issue; otherwise busy accounts look busier than they are. Second, mark records for issues still open at the end of the period. A failure investigation that spans two years is undercounted in both years if you cut it at the year boundary. Accounts with no records at all still need a minimum, such as one review visit a year, or they drop out of the count.
How do you weight accounts by technical complexity?
Multiply each account's recorded hours by a complexity weight. The weight stands in for work the records do not capture, such as preparation, internal escalation and follow-up reading. Records miss this hidden work unevenly: a repeat-order account generates little of it, while an account with an open failure or a new application generates a lot. The table below sets a complexity weight and a planning action for four account tiers. Every value is illustrative, so replace them with figures from your own records.
Account tier | Recorded hours per year | Complexity weight | What to do at this value |
|---|---|---|---|
Repeat orders, rare inquiries | 14 (illustrative) | 1.0 (illustrative) | Cover by scheduled review and inquiry response; group many per engineer |
Routine test work | 46 (illustrative) | 1.2 (illustrative) | Plan test calendar per account; check trends between visits |
Open failures or trials | 118 (illustrative) | 1.5 (illustrative) | Limit to two or three per engineer at a time; set an end date per case |
New applications | 102 (illustrative) | 1.5 (illustrative) | Plan as a project with its own hours, then reclassify after start-up |
The row that drives capacity is open failures or trials. In this illustrative table, one such account needs about 177 weighted hours a year, more than twelve repeat-order accounts combined.
To set your own weights, ask engineers to log hidden work for one month on a sample of accounts in each tier, then compare the logged hours with the recorded hours for that tier. Review the weights once a year, and move an account to another tier as soon as its situation changes.
Technical service engineer account load, worked example (illustrative)
A worked example with illustrative figures shows how heavily account load depends on the account mix. Start with 1,920 working hours a year, the planning figure from the textbook workload example (INFLIBNET Centre, accessed 2026). Take out an illustrative 40% for travel, training, internal meetings and documentation, which leaves 1,152 account hours. Weighted hours per account follow the tier table above: about 14 for repeat-order accounts, 55 for routine test accounts, 177 for open failures or trials and 153 for new applications, all illustrative.
On these illustrative figures, an engineer with only repeat-order accounts could cover about 82 of them, since 1,152 divided by 14 is 82. With only open-failure accounts, the same engineer could cover about 6. A realistic mix of 20 repeat-order accounts, 8 routine test accounts and 2 open-failure accounts uses about 1,074 weighted hours. That fits within 1,152 and gives 30 accounts, with some slack left for new inquiries. Add one new-application account and the total rises to about 1,227 hours, which is over capacity. One repeat-order or routine account should then move to a colleague while the new application runs.
Across these mixes the answer changes by more than ten times. That spread is why a single accounts-per-engineer number misleads.
Which signals show a technical service engineer is carrying too many accounts?
An overloaded technical service engineer shows up in the service records before customers start to complain. Watch five signals for each engineer, and compare them with that engineer's earlier months and with colleagues.
First-response time on technical inquiries rises while inquiry volume stays flat.
The backlog of samples or tests waiting for interpretation grows, and results reach customers after the decision they were meant to inform.
Repeat escalations come from the same engineer's accounts, which often means fixes were rushed and the problem came back.
Scheduled review visits slip, or phone calls replace them on accounts that need on-site checks.
Overtime or weekend work becomes regular rather than seasonal.
One signal on its own can have another cause, such as a single difficult failure case. Two or more together over a quarter point to load. Before you move accounts, check where the time goes. A case stuck waiting for a laboratory or an internal approval looks like overload but needs a process fix; the guide on how to find where technical support cases get stuck gives a method for finding it. After any transfer, re-run the workload formula to confirm the new load.
Where the technical service engineer workload formula does not apply
The workload formula does not apply where the account itself does not drive the hours, or where the hours cannot be counted from records. Four situations fall into this group.
A team that has kept no visit, inquiry or test records for at least a year has nothing to divide. Guesses then produce a false precision, so start logging first.
A launch year, a product recall or a supplier change drives work through all accounts at once, and the past year then predicts the next one poorly.
Engineers who differ widely in experience do not get through the same hours of work at the same speed, and the textbook notes that the workload model assumes each person is equally efficient (INFLIBNET Centre, accessed 2026).
Accounts served jointly with a distributor's technical staff split the work in ways your records may not show, so count only your engineer's share.
In these cases, treat the formula's result as a range and review it after six months of clean records. Lubinpla runs a technical support performance analysis that fixes the period and scope first, then reads your own work records to find bottlenecks and improvement priorities; it is a one-time diagnosis, not a staffing calculator.
Frequently asked questions
How many hours a year does a technical service engineer actually have for accounts?
Fewer than their contracted hours. Subtract leave, public holidays, training, internal meetings, documentation and travel from the working year, using your own time records. The textbook workload example plans 1,920 hours and keeps 55% for selling work (INFLIBNET Centre, accessed 2026), but a technical service team should measure its own share.
Why do published accounts-per-engineer figures not transfer between companies?
Each figure reflects one company's account mix, product complexity, travel distances and definition of an account. A software support team that counts engagements per quarter carries a very different load from a lubricant engineer running used-oil analysis and site trials. The same number stands for different work.
Should technical service engineers be split by region or by industry?
Split by region when travel takes a large share of hours and the accounts are similar. Split by industry when the technical questions differ enough that one engineer cannot answer all of them well. Many teams combine the two: regional coverage for routine accounts, and industry specialists for failures and new applications.