Selling the mouse to the person who bought the keyboard nineteen days ago

A gaming setup is assembled over months, not bought in one order: the monitor comes first, the keyboard a few weeks later, then the mouse, the headset and the pad. The brand sold all five categories and almost never sold two of them to the same person.

Its advertising treated each product as a separate acquisition problem, competing for the same cold search traffic five times over and paying full price each time. Meanwhile the warmest audience it had — people who had bought one of its products and were demonstrably still building — was reached by nothing at all.

We mapped the order and then advertised to it.

• Sequence analysis in Amazon Marketing Cloud, the only place a per-customer order sequence exists — the median gap between first and second purchase per pair came out at nineteen days for keyboard → mouse and about six weeks for monitor → keyboard.
• DSP audiences of the brand's own recent buyers, activated at the gap rather than immediately after purchase.
• Creative that assumes the first product — the ad shows the piece they already own next to the piece they don't.
• Search budget held flat, so the growth could be attributed to the sequence work rather than to more cold traffic.

The second-purchase rate went from 9% to 31%, and revenue per customer rose 22%.

That is what lifted total revenue 29% without any increase in Sponsored spend: the brand stopped paying five separate acquisition costs for the same person and started selling the second, third and fourth item to a customer it had already paid for. The nineteen-day gap is now a scheduled flight rather than a coincidence.