Segmentation

A sentence written for everyone fits nobody. The first ad copy Pagedeck ran was "a document tool for every team." Clicks came in; payments didn't follow.

Splitting the incoming traffic apart showed why. Solo users wanted something cheap that opened fast, a twenty-person team wanted co-editing, and a company of several thousand asked about permission controls and security documentation first. Three groups after three different things, addressed by one sentence, so none of them were persuaded. The line that lands with a solo user is often the one that makes a security reviewer close the tab.

Segmentation is the work of making those three blocks visible. The point isn't the dividing itself but laying out a board you can then choose from.

Start by picking what to divide on. Selling to companies, the usual criteria are size, industry and how the product gets used; selling to individuals, frequency and purpose. Stack criteria and the pieces shatter into fragments, so stop at two. Whatever you pick also has to be something visible in data you already hold; a criterion you can't observe stays a theory.

Three conditions tell you whether the pieces are usable.

1. Are they actually different? If two pieces want the same thing, the split earned nothing. Draw the line where what people want diverges.
2. Are they big enough to count? A piece containing five companies isn't worth building an ad for. A rough test is whether it could carry a campaign of its own.
3. Can you reach them? If there's no way to assemble only those people into an audience, it's a distinction you can't act on.

Estimate the size and revenue potential of each piece. If you already have customers, divide by the record rather than by guesswork — RFM analysis, which grades customers by how recently they bought, how often, and how much, does exactly this.

The result isn't a conclusion. Targeting picks one piece out of it, and positioning then decides what place to occupy in the minds of the people in that piece.

"Split three ways, these all want the same thing. Let's merge them."

Erase a line that doesn't distinguish anything.

"Size crossed with industry gives us twelve pieces. Drop one criterion."

Stacked criteria shattered the market into fragments.

"There's no way to reach this piece. We can't isolate them with ads."

The distinction is real but not actionable.

Pagedeck ended up dividing on size alone: solo, teams of five to fifty, and companies above fifty. Ravi took only the middle piece and rewrote the copy for it, and the same ad budget produced twice the payments. The other two pieces kept running the old copy, which made the comparison easy to read.