Paid Media

Finishing a product teaches one fact quickly: nobody knows about it. The company's own site and blog are spaces it controls completely, and the only people who show up there are the ones who already know the company exists.

The first week's numbers say it plainly. Forty visitors, thirty of them people the team knows personally, and the rest untraceable. Someone hearing about the product for the first time has no reason to open the blog. That person is somewhere else right now, and the company owns nothing there.

Renting somebody else's space is the answer that developed. Money buys a slot where an audience is already gathered. In the era of newspaper pages and broadcast slots, the unit of purchase ran into tens of thousands of dollars, and the only feedback available was whether revenue moved.

Digital advertising changed two things about the arrangement. The entry price dropped to something like $40 a day, and clicks became countable. Both the fact that small companies can advertise at all and the fact that ad spend is now argued about with numbers come from that shift.

Split marketing channels by ownership and there are three, with a trade-off running between them: the more control, the smaller the reach, and the less control, the more the audience believes it. Paid media buys reach with money and pays for it in credibility, because the reader knows they are looking at an ad.

Load a budget, an audience and creative into a platform; slots get allocated by auction, and money leaves per click or per impression.

1. Slots are auctioned. Several companies bid on the same search term, and position is decided by bid combined with a quality score for the creative. That is why the same slot costs a different amount today than yesterday. A competitor raising their budget raises your unit price, which means a channel that worked last quarter can stop working without anything changing on your side.
2. Choose the billing unit. Cost per click charges for people who clicked; cost per thousand impressions charges for times it was shown. Per-click is the default, and the figure tracked for it is CPC.
3. Choose the audience. Targeting by search term reaches people who already know they have the problem; targeting by interest or lookalike reaches people who do not. The first converts better, the second reaches further.
4. Run several pieces of creative. Load three to five variants of headline and image and turn off the weak ones. With only one running, poor results cannot be traced to the creative or the audience, and the usual mistake is changing both at once and learning nothing.
5. Check the arithmetic. At $0.60 a click with 2 percent of visitors signing up, a signup costs $30. If 10 percent of those convert to paid, each paying customer cost $300. That figure is customer acquisition cost, and whether the advertising continues is decided on it.
6. Switch it on and off. Turn it on today and people arrive today; turn it off and it stops the same day. Scaling up raises the unit price, though, because the cheap inventory fills first — a $0.30 click at $40 a day commonly becomes $0.70 at $400.

Four places this channel earns its budget.

1. Message testing — running $40 a day for two weeks right after launch to see which phrasing people respond to. The advertising is functioning as a survey rather than as a traffic source.
2. Search advertising — people who already know the problem arrive at a search box looking for the answer. The highest-converting placement available, and the first thing a small budget goes to.
3. Retargeting — showing the product again to someone who visited and left. Cheap per click and high converting, but the audience is capped by visitor numbers, so it cannot be scaled. A team with 500 visitors a month has a retargeting pool of 500 people and no way to raise it with money.
4. Time-boxed campaigns — a feature launch or a year-end promotion where people have to arrive in one specific week. No other channel can raise its reach on a chosen date.

When budget gets split across channels, what role this one plays is decided as part of the channel mix — advertising handles first contact while the blog answers the questions of somebody already comparing options.

"CPC is at $0.70 and signups are costing $30. Once you follow it through to paid conversion there's nothing left."

Judging on final cost per customer rather than on unit price.

"We doubled the budget and signups went up 1.4 times."

Reporting that the return did not scale with the spend.

"Let's turn ads off this week and look at blog traffic on its own."

Establishing the baseline without paid traffic in it.

Pagedeck started with search advertising when Ravi opened the paid plan. Twelve search terms around document collaboration tools, $40 a day, and after two weeks the numbers were $0.48 a click and 38 signups — about $14 per signup.

That looked good enough to raise the budget to $230 a day. A month later the click price was $0.85 and a signup cost $26, because the cheap search terms had no volume left in them. Ravi pulled the budget back to $90 a day and moved the rest to the blog and the newsletter.

The most common assumption is that switching on advertising produces growth. Money does bring people in, so the conclusion is reasonable on its face. The catch is that the cost of one arrival can exceed what that person will ever pay. Spend $300 to acquire a paying customer who pays $150 a year and doubling the spend doubles the loss. This channel does not create growth; it runs an already-profitable structure faster. Which also means the useful question before increasing spend is about retention and pricing, not about the ads.

The second misreading is that doubling the budget doubles the result, which is an arithmetic anyone would expect. In practice the cheap inventory sells first and only expensive inventory remains. Six times the budget can produce three times the traffic with the unit price doubling on the way. Which is why ad budgets get managed against a target unit cost rather than a target spend, with a rule to stop once that cost is exceeded.