Last Agency
• marketplace
OSTARA wordmarkOSTARA

ostara market.

Doubling Ostara's two-sided GMV in fourteen months while shifting acquisition from paid to organic-driven

The Build | The Run | The Handoff | The Exit

OSTARA wordmarkOSTARA
status

18 months standalone

what was bleeding

Ostara is a marketplace for independent ceramicists. Two-sided. Supply growth had outpaced demand for nine months. Paid was the only demand engine. The founders had stopped opening the weekly dashboard because it depressed them.

months 0–3

the build

Months 0-3 we rebuilt the demand side. Programmatic SEO across artist + city + craft-style — 4,800 pages, indexed in waves. A reworked discovery feed on the site itself. A weekly editorial drop that gave the homepage a reason to be revisited.

months 3–12

the run

Months 3-12 GMV doubled. Organic share of new buyers climbed from 11% to 47%. We added a buyer-side referral that paid out in store credit and broke even by month nine. Paid spend stayed flat in absolute terms while becoming 31% of acquisition instead of 89%.

mid-chapter — the numbers
Blended ROAS climbed from 1.8x to 4.4x over the engagementBlended ROAS climbed from 1.8x to 4.4x over the engagement

Blended ROAS climbed from 1.8x to 4.4x over the engagement

months 12–15

the handoff

Months 12-15 Esme, who'd joined as an editorial assistant in month four, took over the weekly drop and the SEO calendar. We trained a freelance buyer to handle the residual paid. The founders themselves took back the supply side, which had quietly stabilized.

months 15–18

the exit

Last invoice in February. The first programmatic page we built ranks #2 for its phrase eighteen months later. Ostara shipped a print magazine in April we knew nothing about. We bought a copy.

what esme said

What Esme Cavendish said

I'd been on the supply side of the marketplace for so long I'd forgotten there was a demand problem. Last Agency built the demand engine in eleven months and then — this is the part I still can't get over — left. They wrote a 42-page document, made me sit through a Friday where I ran the meeting and they didn't speak, and then sent a goodbye email. They trained the team that replaced them and walked out. I keep recommending them to other founders and watching the same look of confusion cross their faces. Yes. They actually leave.
Founder avatar: ECEC
Esme CavendishCo-founder · Ostara Market · London
results, not prophecies

the numbers that stopped moving when we left.

  • ROAS4.4x blended (paid + organic-attributed)
  • CAC reductionPaid share 89% → 31% of acquisition
  • revenue scale2x GMV in 14 months ($4.2M → $8.7M)
  • time to in-house15 months

more graduates.

• the offer

same shape engagement. your name in this slot next.

Twelve to eighteen months. Fixed scope. We build the function, run it, train your hire, hand over the playbook. Then we're gone.

tell us what's bleedingback to the work