Jordan Gladstone

You can start by making ad creatives using your phone. Even with just selfie videos you can test a ton: different hooks (starting with a question, moving the camera in the first 2 seconds, etc.), different call to actions, sunglasses on or not, etc. - selfie videos, friends, etc.

For creative production Dave organizes shoots to be able to produce hundreds of creatives by having 6 different setups (1 that looks like a kitchen, 1 like a gas station, etc.), different actors (different ethnicities, different ages, etc.) and "cycling people through".

Then start filming other people to be able to show the app more. It helps figuring out which messaging works as well as test different thumbnails (e.g. empty gas), showing or not the app (showing the app works better), actors, locations, etc.

Dave only focuses on evergreen creatives, not on anything seasonal. If it can't be always on and evergreen, they don't do it.

You only need limited resources to be able to respond to FB comments. Do it yourself at the beginning, find unemployed comedy actors and later you can get community managers.

Something that has worked well for Dave is responding to Facebook comments on their ads "as Dave" (they now have a style guide with what Dave likes). It allows for ads to look more organic and creates more engagement, leading to lower CPMs.

Try best practices (e.g. Facebook) but if it's not working, think outside the box. There might be a way that works better for your app.

Before you invest big on TV, do a test with a scrappy creative (with enough quality to still be approved). Dave started with a $250 creative and it actually crushed it in performance (more than some expensive agency ads created later) and they put a lot of money behind it.

Be very flexible during the learning phase with TV, especially if you come from digital and used to understanding attribution perfectly. A CPA on TV is going to look very different from Facebook but you will see huge scale on TV as well as an halo effect on other channels allowing for bigger overall scale.

What's going to take you from 0 to 1M users and from 1M to 3M is probably different things. Some things stop working so you have to constantly force yourself to change and try new things.