How Timehop, running out of money, ditched programmatic ad partners, built its own ad server, boosted CPMs from $2 to $14, and increased revenues 10x YoY
Timehop had three months of runway in the bank when it decided its current advertising solution wasn't going to cut it. Tweets: @digiday , @rickwebb , @bryce , @codinghorror , @digiday , @mdudas , and @digiday Tweets: @digiday : Timehop's 15-person team ended up creating its own ad server. A year and a half later, CPMs have grown from $2 to $28 http://digiday.com/... Rick Webb / @rickwebb : Been a busy and challenging two years but things are pretty sweet right now. http://digiday.com/... Bryce Roberts / @bryce : “The company plans to use this new financing to...” And by “financing” they mean a 1,000% increase in revenue at 50%+ margins. So impressed with the turn around underway at @timehop and the creativity and commitment of the team. http://digiday.com/... Jeff Atwood / @codinghorror : I think @pinboard will enjoy this story; if your company is in crisis, build your own ad server and enjoy success! http://digiday.com/... @digiday : Timehop had three months of runway in the bank when it decided to build its own ad server http://digiday.com/... Mike Dudas / @mdudas : Timehop increased revenue 10x in a year by ditching crappy publisher ad tech and building its own ad server http://digiday.com/... @digiday : “Publishers if they have the control themselves can determine how much they're truly valued” http://digiday.com/...
Context & Ripple Effects
Programmatic advertising has been the default monetization path for small publishers ever since Right Media built the first internet ad exchange in 2005 — intermediaries take a cut of every impression in exchange for automated demand. Timehop's bet ran against that grain: with about three months of runway left, its 15-person team ripped out its programmatic partners and built an ad server in-house.
The payoff reframes what a small media company can do alone: CPMs rose from roughly $2 to $14 (with reported levels up to $28) and revenue grew about 10x year-over-year. The move anticipates a broader pattern of publishers reclaiming their ad stacks, later echoed when the [[a:953864|New York Times committed to dropping third-party data targeting and building its own first-party data platform]].
First-order effects
- Timehop converts near-death economics into a working business: a 15-person team now runs its own ad operations, capturing the intermediary margin that previously went to programmatic partners and lifting CPMs from about $2 to $14.
Second-order effects
- Timehop's former programmatic partners lose that inventory outright, and comparable small publishers face harder questions about why they pay exchange tolls when a tiny team can internalize the stack.
Third-order effects
- If the pattern holds, mid-tail publishing splits between companies that own their ad infrastructure and data — like the Times' first-party platform play — and those renting it, while quality-controlled direct sales push back against the junk-ad dynamics plaguing the open programmatic market.
The trend: Publishers are progressively disintermediating the programmatic layer they helped create, trading outsourced ad tech for owned servers, direct sales, and first-party data.