Blog · Economics
What a reply actually costs, next to every other way to get a lead
Every acquisition channel gets compared on cost per lead, and almost every published number is useless to you — it comes from someone else's category, someone else's price point, someone else's definition of a lead. So this is not a table of benchmarks. It is the arithmetic, so you can compute your own and know what you are looking at.
One warning before the maths: a conversation is not a lead and a lead is not a customer. Channels get compared dishonestly all the time by quietly counting different things. Pick one unit and use it everywhere. Here the unit is a real conversation with someone who has a problem you solve.
The formula
For any channel:
cost per conversation = (money spent + hours spent × what your hour is worth) ÷ conversations started
The second term is the one everyone leaves out, and it is the one that decides the answer for a founder. If you value your hour at nothing, every channel that costs only time looks free — and you will pick the one that quietly eats your week.
Pick a number for your hour and keep it. If you are a founder, a defensible figure is what you would have to pay someone to do the work at the quality you do it. Not your salary — the replacement cost.
Answering on X
By hand, the inputs are simple because there is only one.
- Money: zero, or the cost of a tool.
- Time: running your searches, reading, discarding, writing. In practice twenty to thirty minutes a day once the queries are written.
- Output: however many replies you actually send that get an answer.
So the calculation for a week looks like: (5 days × 0.5 h × your hourly rate) ÷ conversations that went somewhere.
The number that decides everything here is not your hourly rate — it is the denominator. Two founders spending the same half hour get wildly different costs depending on how many of the posts they read were actually buyers. Which is why telling a mention from a buyer is not a nicety: it is the single biggest lever on the cost of this channel.
A rough sense of the shape: if a half hour surfaces two conversations worth having, this channel is very hard to beat. If it surfaces one every three days, you are paying a lot per conversation and should either fix the queries or accept that your market is not here.
What to compare it against
Four honest comparisons, each with the term people forget.
Paid ads
(ad spend + hours managing it × rate) ÷ conversations
The forgotten term is management time, and the forgotten adjustment is that a click is not a conversation. If you count clicks you will always conclude ads are cheap. Count the people you actually spoke to.
Ads win decisively on one axis: you can turn the volume up. No amount of effort makes more people ask for a CRM this week than are asking.
Cold email
(list + sending tools + domains + warm-up + hours) ÷ conversations
The forgotten terms are domain reputation and the slow tax of writing sequences that stop working. Cold email is the closest competitor on price and the furthest away on posture: you are interrupting someone who did not ask, where answering on X means arriving after they did.
It also scales in a way replying does not, which is the honest case for it.
An SDR or an agency
(salary or retainer + tooling + your management time) ÷ conversations
The forgotten term is ramp — the first weeks produce very little, and if you compare month one you will get a number that flatters nothing. Compare a quarter.
This is the expensive option that buys the thing none of the others do: someone else's hours instead of yours.
Content and SEO
(writing hours × rate + tools) ÷ conversations, over a year
The forgotten term is time. Content has no cost per conversation in month one because it has no conversations in month one. Measured over a quarter it looks terrible; measured over two years it is often the cheapest thing anyone does. Any comparison that does not name its window is meaningless.
The comparison that actually matters
Not cost. Posture.
| Channel | Who started it | What that changes |
|---|---|---|
| Ads | You, at a stranger | You pay for attention and hope for need |
| Cold email | You, at a stranger | Interruption. Volume compensates for fit |
| SDR calls | You, at a stranger | Same, with a person absorbing the rejection |
| Content | Them, eventually | They arrive already interested, months later |
| Answering on X | Them, an hour ago | They stated the need in writing, today |
A conversation where the other person wrote down their problem first converts differently from one where you guessed. That is not a marketing claim, it is the difference between answering a question and asking to be listened to. It is also why cost per conversation understates this channel: the conversations are not interchangeable.
Where answering stops being the cheapest option
Four situations, and they are worth being honest about because getting this wrong wastes months.
Your market is not on X. Some categories talk on Reddit, in Slack groups, at conferences, or nowhere public at all. The test is cheap: spend twenty minutes with the operators on a seven-day window. If a broad, correctly built query returns nothing three weeks running, believe it.
You need predictable volume. Demand arrives when it arrives. If you have a number to hit this quarter, this channel is a supplement, not a plan.
Your deal size is very small. Half an hour a day is worth it for a $49/month subscription only if the customer stays. For a one-off small purchase the arithmetic rarely works.
You cannot keep the habit. This is the real one. The channel is cheap in money and expensive in consistency, and the failure mode is never the first week — it is week five, when the searches quietly stop getting run. A channel you abandon has an infinite cost per conversation, because the denominator goes to zero.
Run your own numbers
Do it for one week before deciding anything. Log the minutes, count the conversations that produced a reply, divide. Then do the same for whatever channel you are currently spending on, using the same definition of a conversation.
Most people find the comparison is not close in either direction — either this channel is obviously worth twenty minutes a morning in their category, or it obviously is not. The expensive mistake is running it half-heartedly for six months without ever doing the division.
If you want to see what the volume looks like in your category before spending a week on it, browse who is asking on X by category — real posts, real dates. Or run it on your own offer and count what comes back.
Related: the complete guide to finding buyers on X, and how to reply without sounding like spam — the part that decides your denominator.