Originally published on the Liftoff blog on 20 November 2018, shared here in October 2019. Some platform details have changed since; this is a 2015 to 2016 mobile games story from before iOS 14.5, so the tactics are history, but the order of operations still holds.
Short answer: Shamanth Rao took spend from $0 to $1.5 million a month in 11 months while in-house, as Director of Growth at PuzzleSocial. The order was: fix retention first and spend nothing until it was good, build an LTV model, ramp spend only as daily active users and revenue grew, and let cash decide how fast to go. The original Liftoff page now redirects elsewhere, so this summarises the archived article.
Page Contents
Where did it start?
In spring 2015 Shamanth joined PuzzleSocial, maker of Daily Celebrity Crossword, as Director of Growth. The team was five full-time people plus a couple of contractors. The first recommendation to the team, written on a whiteboard, was “0”: do not spend yet, because retention was not good enough and nobody knew whether ad spend would come back in full.
What did the ramp look like?
- Retention first. Once retention improved, acquisition opened up, still with a rule that spend had to earn its money back, tracked with an LTV model.
- Only high-retention users. That meant saying no to incentivized traffic, cheap traffic and geographies that would not hold up long term.
- Daily checks on DAU and revenue. Spend rose gradually from zero to five figures, then six, then seven figures a month as the curves trended up.
- Cash as a constraint. The team’s LTV model projected revenue over about 12 months, so the team tracked metrics for short and long-term financial sustainability, and cash determined how much they were willing to spend. The cash story is in the world’s best LTV is useless without cash.
How did channels fit in?
The team started with three core channels and ran one or two small experimental ones at a time. When a new channel showed consistently strong performance, they doubled down, which took Liftoff spend to over $200,000 a month. They then widened into every format (video, social, search, playable and display), tailoring the approach to each, and ended with 15 user acquisition partners at $1.5 million a month. Per the article, daily active users grew 7x in under a year, and PuzzleSocial was acquired by Zynga in 2016.
What carries over today?
The playbook (retention before spend, cash-aware ramp, new channels as tests) maps to our current guidance: see how to scale a Meta ads budget for app campaigns and the choice between in-house, agency and freelance support.
Frequently asked questions
Was this done in-house? Yes. It was done by an in-house team at PuzzleSocial, five full-time people plus a couple of contractors.
