A full-time person needs about forty hours a week of work. Not forty hours of tasks you would like to be rid of: forty hours of work that can be described and checked, which is a narrower category and the one that decides whether this succeeds.
The measurement
For two weeks, log what you and the relevant people actually do, in half-hour blocks, marking each block against two questions. Could this be written down clearly enough for a competent stranger to do it? Could somebody tell afterwards whether it was done correctly?
Blocks that answer yes to both are candidates. Everything else is not, whatever it feels like.
It takes about twenty minutes a day and almost every business that does it is surprised in the same direction.
The usual result
Between ten and twenty hours a week, against an intuition of forty or more. The gap comes from three places: work that felt routine turns out to require judgement, work that felt substantial turns out to be occasional, and interruptions turn out to occupy more of the day than anybody believed.
What to do when it is short
Combine roles. Two half-roles for one person, with a clear priority between them. This works and it has a specific failure mode: a role assembled from everything left over becomes a role nobody can define, and it produces the worst outcomes in this industry.
Two related halves work. Five unrelated quarters do not.
Share across functions. One person supporting two teams, with a named person deciding priority when both want the same afternoon. Without that named person it degrades into whoever asks loudest.
Wait. Underrated, and frequently correct. The cost of waiting six months is close to nothing; the cost of hiring into fifteen hours of work is an engagement that ends badly and a business that concludes offshoring does not work.
The growth assumption
The most common reason a short role is hired anyway is that the volume is expected to grow into it. Sometimes it does.
When it does not, the person spends months underused, which is worse for them than for you. Underused staff disengage, invent work to look busy, and leave, and the leaving is usually blamed on something else.
If the case depends on growth, say so explicitly, put a date on it, and agree what happens if the growth has not arrived.
Bursts
Work that arrives in peaks and troughs suits a full-time hire badly wherever that hire sits. Month-end, seasonal trade, campaign cycles: a person sized for the peak is idle in the trough and a person sized for the trough cannot cover the peak.
The honest answers are overtime, a shared resource, or accepting that the peak is uncomfortable. Sizing a permanent role to a peak is how businesses acquire capacity they resent paying for in March.
The part-time question
Part-time offshore arrangements exist and are less common than clients expect, because a part-time role is harder to recruit for and harder to retain: the person is looking for full-time work and will take it when it appears.
Where a provider offers part-time cheerfully, ask about tenure in those roles specifically. The number is usually different from the full-time number.
The threshold, stated plainly
Thirty hours a week of describable checkable work is enough to proceed with a full-time hire, because the remainder fills quickly once somebody capable is looking for it. Fifteen is not, and no amount of enthusiasm changes that.
Anybody who tells you otherwise before seeing your log is selling.
What underused staff actually do
Not nothing. They fill the time, because sitting visibly idle in an office is uncomfortable. They take on work outside the role, they elaborate the work they have, and they build processes more complicated than the task requires.
That output looks like productivity and is worse than idleness, because it creates things somebody now has to maintain. When the engagement is reviewed, the team appears busy and the value is hard to locate.
The utilisation question to ask a provider
What happens if the role turns out to be twenty-five hours. Some providers will say the person can support another client, which is a different arrangement from the one you thought you were buying and should be agreed rather than discovered.
Others will say the person is yours regardless, which is honest and means you are paying for idle time. Both are workable and only one of them is usually explained.
Measuring again later
Repeat the two-week log a year in. Roles drift, and the work a person is actually doing after twelve months frequently bears limited resemblance to the role that was recruited for.
That drift is not necessarily bad. It is worth knowing about, because the role description, the recruitment criteria and the replacement plan all still describe the original.
The log, in practice
A shared sheet with three columns: the half-hour, what happened, and two yes-or-no marks. Nobody enjoys it and two weeks is short enough that people finish it.
Do it for everybody whose work might move, not only the person you have in mind, because the candidate hours are frequently spread across several people rather than concentrated in one.