Most complaints about connects are really complaints about a hidden price becoming visible. Bidding was always expensive. It cost time, and a bad proposal cost exactly as much time as a good one. Connects did not make bidding costly; they made the cost countable.
Which is genuinely useful, because a countable cost can be managed and a vague one cannot.
Three numbers
Open your account and write these down for last month. It takes about ten minutes and almost nobody has done it.
- Connects spent. Including any you bought.
- Proposals sent. This number on its own means nothing, which is the point.
- Proposals that got any reply. A rejection counts. A question counts. Silence does not.
Now divide the first by the third. That is your cost per conversation, and it is the only price you are actually paying.
The number is usually a surprise
Freelancers who run this for the first time typically find that most of the month's connects went on one category of job they keep bidding on out of habit, and that the jobs that became work came from a much smaller and more obvious set.
Why volume is the wrong lever
If twenty proposals produce one reply, forty produce two. You have paid twice as much to stay in the same position, and you have spent twice as long doing it.
Reply rate is not something that improves by repetition. It is a property of what you write and which jobs you write it for. Doubling volume changes neither.
What to stop bidding on
The fastest way to cut cost per conversation is to stop paying for proposals that were never going to get one. A useful filter, applied before you spend anything:
| Signal | What it usually means |
|---|---|
| No budget and a two-line description | The client has not decided what they want yet |
| Client has hired nobody, ever | A coin toss you are paying to enter |
| Forty proposals within the hour | You are betting on being opened by someone who has stopped reading |
| Nothing specific enough to open with | You cannot do the one thing that wins proposals |
| You are a defensible choice, not an obvious one | Defensible loses |
Write the rule down. A filter you have to re-decide each time is not a filter, and it will not survive a slow week, which is the only week it matters.
The experiment worth running
Halve your volume for two weeks. Put the time you get back into reading the postings properly and writing openings from scratch.
One of two things happens. Conversations hold steady, in which case you have just halved your costs and freed up hours. Or they drop, in which case you have learned that your volume was doing real work and your filter is too tight. Both of those are better than not knowing.
The longer version of this arithmetic, including when boosting is worth it, is in connects and bidding math.
The way out
Every proposal costs connects. An invitation does not. That difference compounds quietly, and it is why the freelancers who look like they are barely trying are usually the ones who stopped bidding some time ago.