$580K
Blended revenue
+82% YoY
The strongest commercial month of the year so far, and a checkpoint on the question we raised in April: not whether we can win customers, but whether we can keep them.
In April we drew a line through the funnel. Winning new customers was going well. Keeping them was not.
We said that second half, the part after the first purchase, was where the next stage of growth would be decided. We called May and June the window to start closing the gap.
This review picks up exactly there. May was a strong month for sales, so the top line is not really the question. What matters is what is happening underneath it. The report walks through that in three steps, then looks at each channel, the audience, and what we do next.
Each question has a clear answer this month. Two are good. One has not changed yet.
Comfortably. May brought in 1,168 new customers, up 54% on last year, supported by a 123% rise in ad spend. Traffic was very high. There is no sign of a demand ceiling.
Less clearly. The cost to win a new customer has risen for three months: $110 in January, $150 in April, $158 in May. Conversion eased to 0.77% even as traffic climbed. This is still manageable for now. It is the direction to watch, not the level.
This is the one that matters, and it has barely moved. The returning rate went from 30% to 31%. For a brand at this price point, a healthy figure sits closer to 35 to 40%, so there is real distance to cover. While we acquire at full pace without lifting retention, much of the spend is replacing customers rather than adding them.
Google and Meta carried most of the spend. The clearest change this month was on Amazon, which has its own chapter next.
$130,909 in spend returned $529,637 in tracked value, a 4.05 times return. Performance Max did most of the work at 4.31 times. Branded search was the most efficient at 4.89 times. Non-Brand search was the weakest at 2.80 times, and is the clearest place to tighten.
$50,120 across a four stage funnel. Awareness reached 1.39 million people, consideration drove 63,982 landing page views, and the two conversion campaigns delivered 468 purchases. One issue to fix: retargeting showed the same people ads 19 times on average against a pool of only 30,697, so that audience is saturated and needs widening.
After the March test, Reddit was effectively off in May, with almost no spend and only 36 sessions. Nothing to report this month. Whether to restart it in Q2 is a decision for us to make.
Amazon was the story of the month. After a difficult spring, it delivered its strongest sales month since the relaunch began last September.
Revenue of $38,503 was up 32% on last year and more than double April. Advertising returned 4.65 times spend, the highest the account has recorded, at a 21.5% cost of sale that sits well inside the under 30% target the client has set. Ad-attributed sales of $25,746 made up about two thirds of the total.
Two things. First, the Buy Box was restored to 82%, up from 33% a year ago, so the listings were winning the sale again. Second, the campaign structure did its job. Sponsored Products Auto was the engine, turning $2,042 of spend into $16,637 of sales at 8.15 times, the best single campaign in the account's history. The branded campaigns held above 4 times alongside it.
Units actually fell 29% year on year, to 131. The gain came entirely from price: average selling price reached $303, up 77%. The higher priced Stormcloud and Driftwood colourways are being accepted by the market, led by the Headrest XL and Large formats. That is a healthier kind of growth, but it means volume is still something to rebuild.
The Buy Box dropped briefly twice, over the first three days of May and again on the 19th and 20th. Both recovered within 48 hours, and neither needed the heavy spend throttling that hurt April. That is the clearest sign that the process fixes we put in last month are holding.
The constraint is no longer account health, it is the top of the funnel. Page views were down 49% year on year, and the listing converts at 0.82% against a category benchmark of 2 to 3%. The account converts well when people arrive; the job now is getting more of them there.
Amazon matters beyond its own numbers. The April fixes there were specific and operational, and they produced a record month in about four weeks. That is the template we want to apply to retention next.
The audience is broad and established. Of the visitors we can identify, the age bands from 25 to 64 sit close together, with 55 to 64 now the largest single group. It is national, led by California at 17% of users, then New York, Texas and Florida, and it continues to pay a premium, with average order value holding at $299.
On the product side, the Original Orthopedic Bed still drives 63% of revenue. The notable shift is seasonal: the BarkerChill cooling range and the travel beds both grew as the weather warmed, a useful tailwind into summer.
If May leaves one thing to act on, it is retention. We have shown we can win customers at scale. The lever that changes the economics from here is keeping them.
Amazon shows how that can work. The fixes there were operational and specific, and they turned a falling channel into a record month in about four weeks. The returning customer problem can be approached the same way, as a system to build rather than an outcome to hope for.
The plain read of May: a strong month for sales, with acquisition working and Amazon recovered, but retention up only one point in the month we said it needed to move, and acquisition slowly getting more expensive. The half of the funnel after the first purchase is still where the next stage is decided.
May was the strongest commercial month of the year. The half of the funnel after the first purchase is still where June will be decided.