Every Upwork job page carries a small block of numbers about the client: how much they have spent, how many people they have hired, how many proposals the job already has. Most freelancers glance at the budget and skip the rest.
That block is the closest thing you get to a reference check, and reading it properly is the cheapest skill in freelancing. Connects cost real money, and the difference between a good week and a wasted one is usually not the quality of your writing. It is which jobs you wrote for.
Hire rate is the one that matters most
A client who has posted ten jobs and hired on nine is telling you something reliable: they post when they intend to buy. A client who has posted ten and hired on one is telling you something equally reliable, and it is not about you.
Jobs from low-hire-rate clients are not lost on merit. Frequently they are never awarded at all — the budget evaporated, the internal hire happened, the project got shelved. You can write the best proposal of your career and lose to nobody.
This is the single most useful number on the page, and it is also the one most freelancers never look at.
Proposal count is about timing, not quality
A job with fifty proposals is not necessarily a bad job. It is a job you are arriving late to, and the client is now skim-reading rather than reading.
The practical effect is that your first line has to work harder, and any proposal that opens by describing you rather than their problem is going straight past. Under five proposals is a different situation entirely — you are early enough to be read properly, which is worth more than most people realise.
Treat the number as a difficulty setting rather than a verdict.
Interviews in progress is a stop signal
If the client is already interviewing several people, the shortlist exists and you are not on it. Applying is not impossible — shortlists collapse — but you are now spending connects on a longshot rather than an opening.
This is the number to check last, because it can override everything else. A perfect-fit job with five people already in interviews is usually a worse bet than an average job posted an hour ago.
Invites sent tells you how they hire
A client who sends a lot of invites is hunting rather than waiting. That cuts both ways: they are engaged, which is good, but they may already have a preferred candidate they invited directly, and the public posting is a formality.
Combined with a high interview count, heavy invite activity usually means the decision is close to made.
Total spend is context, not quality
A high total spend tells you the client is real and has paid people before. It does not tell you they pay well, and it certainly does not tell you they are pleasant to work with.
A low spend on a new account is not automatically a red flag either — everyone has a first job. It just means you have less to go on, so the other numbers carry more weight.
Putting them together
No single number decides it. A great client on a crowded job and a mediocre client on a quiet one can be equally worth applying to, for opposite reasons, and that is exactly the judgement that is hard to make forty times a week when you are tired.
That is what our free Upwork job scorer does. You paste in the numbers from the job page and it weighs those five factors — hire rate, competition, client history, interviews in progress and invites sent — into a single score out of ten with an apply, maybe or skip verdict. It is arithmetic rather than a language model, so it answers instantly and gives you the same answer twice. It needs no account.
The point is not that the score is always right. It is that a deliberate decision, made in ten seconds before you write anything, beats the decision you make halfway through writing when you have already sunk twenty minutes into it.
What to do with a "maybe"
The mixed cases are the interesting ones, and they usually mean the signals disagree — a strong client on a crowded job, or a quiet job from a client with no track record.
The honest rule: apply if you can say something specific in the first two lines, and skip if you cannot. A maybe job rewards a proposal that proves you read it and punishes a generic one harder than an easy job would.
If you are not sure what specific looks like, we have worked cover letter examples that show the structure — open with their problem, one piece of proof, one question.
The connects arithmetic nobody does
Work out what you actually spend. If connects cost you real money and you apply to thirty jobs a month, that is a line item, and the return on it is entirely determined by which thirty you picked.
Cutting ten bad applications is worth more than improving all thirty proposals by a small margin, and it takes less effort. Most freelancers do the opposite: they optimise the writing and never touch the selection.
Read the numbers first. Then write.
