guides, updates, and news from the numu team.
the generic AI look is a grounding problem, not a model problem.
AI can ship a hundred ad creatives a week and most will lose. three myths about AI ad creative for mobile games — and the loop that turns it into installs.
mobile game UA benchmarks 2026: CPI, IPM and ROAS by genre.
what 'good' looks like by genre and platform — and a three-lever framework for telling whether a soft number is your creative, your audience, or your product.
introducing numu: the growth engine that doesn't sleep.
we spent a decade running UA at garena and hit the same ceiling every studio hits — you can only hire so many senior people. numu removes the ceiling. here's what it does today.
how numu cut CPI 30% on a live RPG in six weeks.
a real campaign, start to finish: what numu saw that the team couldn't, what it changed, what it left alone — and the number at the end.
mobile game user acquisition in 2026: an operator's playbook.
the UA analyst isn't dead — the job moved. the five levers that actually move installs now, and why the edge shifted from spreadsheets to loop speed.
creative fatigue on facebook ads: the 4 signals frequency won't show you.
by the time frequency spikes, the CPI damage is already booked. the four leading signals numu watches instead, the thresholds we use, and a 10-minute weekly audit.
what meta's 2026 changes mean for game studios.
advantage+ is the default, IPM signal is noisier, and the old bidding heuristics overshoot. our read from live accounts — and what to change monday.