Waiting for Marta to reply
A 6,000-word medical device manual needs translating into Italian by Thursday. Your project manager's first choice is Marta, who translated the last manual for this client and knows the terminology. The PM emails her at ten. At eleven there is no reply. At twelve the PM emails a second translator, who replies that she is busy. At two Marta replies from a train: she can do it, but only by Friday.
Now the PM has to decide between a translator who knows the client and one who does not, while a second job for the same afternoon has not even been started.
Why assigning jobs takes so long
- Translator knowledge sits in PMs' heads: who is good at medical, who knows which client, who is reliable on deadlines.
- Availability is only discovered by asking.
- Offers go out one at a time, because sending to several at once risks two people accepting.
- Key details, such as word count, deadline, rate and file type, are written into each email by hand.
- The vendor database, if there is one, is out of date on rates and specialisms.
This is the core of a PM's day at many agencies, and most of it is waiting.
What the delays cost
Jobs start late, which squeezes translation and revision time. Clients' preferred translators are not used as often as they could be, which affects consistency. PMs spend hours on email ping-pong and cannot handle more projects. Good freelancers get irritated by offers that arrive late or are withdrawn after they said yes, and they drift to agencies that are easier to work with.
| Step | By email | With matching and offers |
|---|---|---|
| Choosing who to ask | Memory | Shortlist by pair, specialism, client history |
| Checking availability | Ask and wait | Stated availability shown, offers to several |
| Sending details | Typed per email | Standard offer with word count, deadline, rate |
| Acceptance | Email reply | One click, first suitable acceptance wins |
| Declines | Silence or a reply | Recorded, next offer goes automatically |
How we build translator matching
- Your vendor records hold each freelancer's language pairs, specialisms, CAT tools, rates, and which clients they have worked for, with quality notes from your reviewers.
- Translators mark their availability and capacity through a simple page, such as 'available this week, up to 3,000 words a day', or block out holidays.
- When a job is created in your system, a shortlist is produced: matching pair and specialism, clients' preferred translators first, then others by reviewer scores and availability.
- The PM sends offers to one, several, or the whole shortlist in waves, with word count, deadline, rate, subject, tool and a preview of the file.
- Translators accept or decline from the email or a phone-friendly page. The first suitable acceptance can be assigned automatically, or the PM chooses from those who accepted.
- Unsuccessful acceptances are told straight away that the job has gone, so nobody starts work on it.
- The job and its files are then released to the translator in your CAT tool or TMS, such as Phrase, XTM or memoQ, and a purchase order is raised.
If you already use a TMS such as Plunet or XTRF, much of this can be switched on or extended in it. We look at that before suggesting anything custom.
What PMs get back
Jobs are assigned without the hours of waiting. Preferred translators are asked first and systematically. PMs can run more projects, and spend their time on problems rather than waiting for replies. Freelancers see clear offers with everything they need to decide, and an answer quickly when they accept, which makes your agency a place they want to work with.
Is this how your PMs spend the day?
- PMs email translators one at a time and wait.
- Knowledge of who is good at what lives in people's heads.
- Your vendor database is out of date on specialisms and rates.
- Two translators have accepted the same job.
- Jobs regularly start late because assignment took so long.