Paid Acquisition

New users used to be counted as one lump. Three thousand signups this month was the whole number, and whether the month went well got argued from that one figure. There was no column separating the people an ad budget brought in from the people who found the product on their own.

The gap showed up when it was time to plan the next month. Grace would ask whether a year target was reachable given three thousand last month, and Ravi couldn't answer. Of those three thousand, 2,200 came from 20 million won of ad spend. Hold the budget and next month looks similar; cut it in half and the number falls by more than half. Two identical counts of three thousand leave very different residue when the spending stops, and counting them as one lump hides that difference.

So teams started splitting arrivals by whether money was spent to produce them. The half made with ad spend is paid acquisition. The half that arrives through search or word of mouth with no ads behind it is organic acquisition. Once the two are separated, the question changes. It stops being how many people arrived and becomes what one person cost and whether that person returns more than that.

You pay for a slot where people are looking, and some fraction of the people who see that slot end up in the product. The top of a search results page, the space between posts in a feed, and the seconds before a video all get sold this way.

The cycle runs in five steps.

1. Decide what counts as a result. Signup, install, or payment — pin one down first, or an ad that only grew its click count gets read as a good ad.
2. Pick the slot. Reaching someone already searching for the problem and showing something to someone who doesn't know they have the problem call for different places.
3. Decide what to bid. Ad slots mostly sell by auction, so when several advertisers want the same slot the higher bid gets shown, and the unit price climbs as competition does. Prices are usually quoted per click or per thousand impressions.
4. Build what gets shown. That means the ad copy plus the landing page the click arrives on — a standalone page built to drive one action. Promise team collaboration in the ad and drop people onto a pricing list and half of them leave on the spot.
5. Measure and adjust. Attribution connects which ad a person came from to whether they actually signed up. Attribution is the method for deciding which touchpoint gets credit for a conversion, and without it there is no basis for choosing where to put more money.

Two numbers report the result. One is what a single new customer cost, the customer acquisition cost (CAC): 10 million won of spend producing 400 signups is 25,000 won each. The other is revenue against spend, return on ad spend (ROAS): 10 million won in and 30 million won out is 300 percent.

Those two aren't enough by themselves, so they get compared against the total a customer leaves behind, the lifetime value (LTV). If someone acquired for 25,000 won only ever pays 18,000 won, the better the ads run the more money is lost.

It buys time with money. The situations where it fits are the ones where turning the budget on today and having people arrive today is exactly the property you need.

The first is testing demand. Before the product is built, put up a single page describing it and run ads for three days. A thousand visitors producing four sign-up forms means that message has no demand behind it, and learning that is far cheaper than learning it after six months of building.

The second is needing people on a fixed date. Two hundred seats to fill for next week's webinar leaves no time to wait for a search ranking, so the budget buys the date.

The third is the early stretch when nothing has accumulated yet. A blog post takes three or four months to reach the top of search results, and with no users in the meantime there is nothing to base product fixes on. Ads put people in now while the compounding channels get built.

The fourth is scaling something that already works. Once it's confirmed that a person costs 20,000 won and returns 120,000 won, the remaining job is to raise the budget. There is a ceiling, though. Channel saturation is the state where more money in the same channel stops improving efficiency: a channel that brought 500 people for 10 million won and brings 700 for 20 million returned 1.4 times the result for twice the money.

Running the other way, while a person returns less than they cost to bring in, a larger budget only widens the loss. What needs fixing there is the product or the pricing, not the ads.

"Sixty percent of this month's signups are paid. Cut the budget and that share goes with it."

A warning that the current growth is riding on spend, and turning the ads off returns the number to where it started.

"Don't just push paid up, look at organic alongside it. Right now nothing is left if we stop advertising."

A proposal to balance the traffic bought with money against the traffic that accumulates by itself.

"Our cost per paid signup went from 20,000 won to 32,000. I think we need to open another channel."

The channel in use has met nearly everyone it can meet.

At Pagedeck the phrase comes up in the monthly metrics review. Ravi puts up arrivals and unit cost by channel, and the first thing Grace asks is not the total signup count but the share of it that carries ad spend. If that share moved from 40 percent last month to 60 percent, the extra signups are the result of spending rather than growth.

There are two choices from there. Leave the budget where it is, or pull part of it across to the side that accumulates.

The most common belief is that money keeps producing more indefinitely. There genuinely is a stretch where doubling the budget doubles the arrivals, and living through only that stretch leads straight to treating budget as growth. But ad inventory is finite, and the unit price starts climbing around the fifth time the same person sees the same ad. Raise the budget without knowing that and six months later arrivals are up 1.3 times while costs have doubled.

The opposite mistake is treating the whole approach as unhealthy, since growing without ad spend looks sounder. But this is a faucet: open it and water comes, close it and it stops. The problem isn't using the faucet, it's not building a bucket to catch the water. Keep widening the faucet while the bucket leaks and the moment the budget stops, six months of arrivals disappear with it.