Cheapest First-Customer Channel: Test Before ChatGPT Ads

You have a list of names and a half-built product. The hard part is finding out who will pay before you spend money you cannot recover. Before incorporation, your first marketing dollar should buy learning, not scale.

The cheap channel is the one you can afford to be wrong on

Start with the channel that costs the least to test, not the one that looks most impressive. Email, informal networking, and a small paid experiment are all candidates, but only one should get your first real spend. Customer acquisition cost by channel is the number that matters, because it tells you what a yes costs before you scale the channel.

A 2026 channel study looked at spend and acquisition results for 214 companies, from January 6 through August 29, 2026. In that dataset, every established paid channel cost more in 2026 than in 2024, with increases from 6.8% to 31.2%. That makes paid ads a later test, not the default. You want to know whether a buyer exists before you pay for attention in a market that is already getting more expensive.

OpenAI opened ChatGPT Ads, a self-serve advertising platform, to advertisers on July 22, 2026. The ChatGPT Ads CAC figures came from live campaign data spanning seven weeks and 14 accounts. The spend-weighted average was $468, higher than the simple average of the seven weekly figures. That number is useful, but it is not a green light for a pre-revenue founder with no finished product.

Warm paths still close faster and cheaper

Referred leads convert at 30% to 45%, while outbound leads convert at 5% to 10%. Inbound demo requests land at 18% to 25%. Mid-market deals take 71 days to close on average, 48 days for referred leads, and 118 days for cold outbound. If your buyer list is small, that gap is your advantage. You are not trying to outbid everyone; you are trying to reach the people who already know you or someone who does.

That is why the first test should be a conversation, not a campaign. Send an email to a named buyer and ask for a short call. Ask a warm contact for introductions. If you get no reply, change the offer, the message, or the channel. Do not add budget until you have a reason to believe the buyer exists.

The checklist keeps the first test small

The checklist below is the whole system. It keeps you from turning a first-customer search into a paid-ads habit before you have evidence. It also gives you a way to say no to a channel without feeling like you are giving up.

1. Write the money you would spend before your first yes for each channel.
2. Cross out any channel you cannot afford to run for a month.
3. Write your minimum-viable-spend floor before spending.
4. Mark the channel to stop or switch when cost to first customer crosses the floor.
5. Put the new ad platform after the cheapest channel.

Ranking channels by cost to the first customer makes the comparison honest. Crossing out unaffordable channels keeps the test small enough that a bad week does not become a crisis. Writing the floor in advance gives you a line you can see, so the decision is not emotional. Marking the stop rule protects you from adding money to a channel that is not working. Putting the new ad platform last leaves the newest option where it belongs: a later experiment, not the first bet.

Set the floor before you spend, not after. Write the amount you can lose without changing your life. Say your floor is $3,000 a month, and cross out any channel that needs more to produce a yes. If it produces a yes under the floor, you have a customer and a number you can trust.

When you do test the new ad platform, keep the same rule. Start with a small amount, watch the cost to a first customer, and stop when the floor is crossed. The platform may be useful later, but it does not replace the cheap test that tells you whether the buyer exists.

Send the first email today, then let the reply decide whether the next step is another conversation or a paid test.