eBay Social

eBay Social launched in Russia in July 2014 as eBay Marketplaces GmbH’s bet on turning enthusiastic shoppers into a distribution channel. The core loop: build a themed product collection (Pinterest-style), share it to VK or Odnoklassniki, and earn 6% cashback when followers buy ⁠— with 3% back to the buyer too. Balance converted into eBay discount coupons ($1 / $3 / $5 / $10, priced off the cashback account) and 15% partner coupons. eBay claimed ~1 million active Russian users at launch (Lenta.ru, Jul 2014). The pilot saw ~500 new users/day and ~100 collections/day. The concept was compelling enough that eBay replicated it in Mexico two years later.

The loop: editorial collections drove sharing, sharing drove purchases, purchases paid cashback, cashback became coupons that pulled users back in

Sharing was the engine: one click sent a collection to VK, Odnoklassniki, Facebook or Twitter

From a collection straight to the eBay listing ⁠— the click that turned browsing into a purchase and cashback

I was the editor of the women’s section, working against a structured brief of five overlapping remits. Each produced a published 40-item collection: affordable outerwear (≤ 999 ₽, autumn/winter), mid-range branded items (1,000–⁠5,000 ₽, harder-to-find in Russia or markedly cheaper than local prices), deals and sale monitoring (e.g. Ralph Lauren women’s discounts, Surfdome, Footasylum Outlet), trend-led inspiration (gadgets, fashion, home, beauty ⁠— sourced from Gonzomag, Engadget, Kosmetista), and open-brief homepage ownership. Each pick required a delivery check, seller-rating check, and written rationale.

Editorial here wasn’t writing ⁠— it was merchandising: a price ceiling, a quality bar, a delivery-and-seller-rating check, and a trend sense, repeated daily across themes

The collections feed: activity trackers, retro Casio, kitchen gadgets ⁠— several of these sets were mine

Doing it at volume surfaced the product’s weak points in a way no brief could. The site needed reliable cashback crediting and autosave for collections ⁠— notes I made directly on the assignment while working ⁠— which became the two issues I escalated hardest to the team.

Editorial standards built into the product: what a review needed to reach Editor’s Choice, and the nudge that told its author what to fix

On 4 July 2014 ⁠— during the product’s pilot, two weeks before the public press announcement on 18 July ⁠— I created the thread “EBay Social ⁠— скидки и акции” on Taker.im, then Russia’s largest independent eBay buyer community. It opened with a structured FAQ explaining the service and became the canonical place users went to understand the product, troubleshoot cashback, and share collections. It grew past 460 pages over the years ⁠— an organic support-and-acquisition channel that cost nothing and ran on its own momentum.

Thread created 4 Jul 2014 ⁠— two weeks before the public press launch

Grew past 460 pages over the years ⁠— the canonical eBay Social reference thread

Beyond Taker.im I also created and administered the product’s VK group (vk.com/ebaysocialru) ⁠— eBay Social’s presence on Russia’s largest social network, where collections and coupon promotions were shared and the embedded collections app lived.

By late 2015 the forum ran its own contest around eBay Social collections ⁠— the community had started organising itself

Every new user who searched ‘ebaysocial’ and landed on that thread was being onboarded by content I’d seeded ⁠— for free, for years

Editing daily and running the community made me the heaviest power-user ⁠— which is where the concrete product feedback came from, not from a brief. Four issues I raised to the team; autosave, search and the CAPTCHA step were reworked and shipped, along with image selection on publish:

Autosave / drafts: collections weren’t reliably saved while being built. A 40-item collection is real work; losing it mid-build kills editorial throughput and discourages users from even starting. Cashback crediting: balance didn’t always appear reliably ⁠— and this was also the #1 complaint in the 460-page community thread. The entire incentive depends on the money actually showing up. Discovery and search: tag- and collection-based navigation was weak. Collections are only fuel if they’re findable after they fall off the homepage. Coupon flow and registration: friction in the CAPTCHA, sign-in, and coupon-activation steps. Every step between ‘I want this’ and ‘I bought it’ leaks conversion.

Sign-in sat between wanting an item and buying it ⁠— a social-network picker over the feed

Signing in with eBay inside the editor ⁠— one more step on the way to a coupon

The four friction points I escalated ⁠— all in the critical path between collection and conversion

The most valuable product feedback came not from a brief but from doing the user’s job at volume

Editing the women’s section daily and running the community made me the product’s heaviest power-user, so I could have escalated almost anything. I spent that access on two unglamorous things: reliable cashback crediting and autosave for collections. Neither is a feature anyone demos. In an incentive product the incentive has to be boringly reliable, or the community says so loudly and in public, and running a thread that reached 460 pages I would have been the first to see it.

The other call was timing. I seeded the flagship Taker.im thread on 4 July 2014, during the pilot and two weeks before the public press announcement, rather than waiting for launch collateral. It opened with a structured FAQ and became the canonical place users went to understand the product: an acquisition and support channel that cost nothing and kept compounding for years.

What I would do differently is all measurement. I never quantified editorial impact per collection in views or conversions, never closed the loop on the autosave and cashback fixes with a before and after, and never instrumented what the thread actually contributed to acquisition. The work was right. The evidence for it was left on the table.

The product reached genuine scale ⁠— 150,000+ test-period users, ~1M claimed at launch, SEO gains of +20% Yandex top-3 / +15% Google top-3 later in its life. eBay took the model to Mexico in February 2016 (ebaysocial.mx), localising collections, reviews and cashback to Spanish ⁠— the Russian product I worked on became the template.

The community thread: 460+ pages is the clearest evidence of what works when you seed a support channel on day one. The thread deflected product questions, onboarded new users, and generated referrals for four-plus years at zero ongoing cost. I’d measure the support deflection and referral value properly now ⁠— I didn’t track it at the time.

What I’d do differently: quantify editorial impact per collection (views, conversions), close the loop on the autosave and cashback fixes to show the before/after, and instrument the community thread’s acquisition contribution. The product lesson is blunt: in an incentive product, the incentive must be boringly reliable or the community will say so ⁠— loudly, in a 460-page thread.

In an incentive product, the incentive has to be boringly reliable, or the community will say so ⁠— loudly