Please don’t copy this setup.
Maximum Delivery is one of the fastest ways to burn your budget.
TikTok has no problem spending every dollar you give it.
Use manual bidding instead.
If your creatives are consistently getting less than 1.2–1.5% CTR, don’t keep increasing the budget.
Test new creatives.
New hooks.
New angles.
New edits.
The biggest wins usually come from better creatives, not bigger budgets.
If you’re building your first product…
Read this post.
One of the biggest mistakes founders make is trying to perfect everything before launching.
More features.
Better design.
More polish.
You don’t need that at the beginning.
Test demand.
Test your creatives.
Don’t spend weeks trying to make the “perfect” ad either.
The market will tell you what needs improving much faster than your own assumptions.
If the market doesn’t care…
No amount of polish will save it.
The reason why my personal site has just one font size, monchrome colors, and looks “boring” is simply because I don’t consider myself a great designer.
By using as little variation as possible, I limit the amount of design mistakes I can make, and instead, I try to make the content itself (hopefully) interesting.
I know a few tricks to not make my work look like shit, but it’s far from the skills of a designer working at Linear for example.
To expand on that, from what I’ve seen, most “design engineers” excel at either design or code, not both. I never believed I can be exceptional at both, and I haven’t seen many people that are really, really great at both.
To me, trying to be great at both is like 2-in-1 shampoo conditioner, it never really works as well as two separate products designed specifically to do one thing well.
That’s why I made open source projects like Sonner and Vaul, that require some design sense, but these are far from sharing design kits for example.
And again, there are exceptions, but you shouldn’t expect all design engineers to design and build absolutely amazing things on their own.
I think most design engineers (including me) are just engineers that care about design, but that doesn’t necessarily mean they are good at designing (and vice versa).
To me, the title design engineer is misleading in that it’s more of a description of where one’s interests overlap instead of actually indicating high skill in both disciplines.
TikTok just rolled out the MCP Server and there are already 29 AI Skills in the marketplace.
Pretty interesting move, gonna test it out.
If it actually works as expected, this could seriously speed up ad setup and campaign launches.
Most of the time it’s clients who come in asking to expand into a new country
I’ll usually do some research first since what works in one market doesn’t always translate to another same goes for creatives, they can hit very differently depending on the region
At the end of the day it really depends on the product. If they’re set on it, we just start testing and see what sticks
You can usually get a rough signal within the first $200–$1k spend
Don't underestimate Tier-3 traffic.
These are some TikTok Ads campaigns I ran for a client app.
A few numbers that stood out:
• CTR up to 3.25%
• CPI as low as $0.01
• Thousands of installs from creative testing alone
Before anyone asks...
No, this wasn't magic.
These results came from testing a lot of creatives.
Different hooks.
Different edits.
Different angles.
Over and over again.
A few important notes:
• It was a client app, so I don't know the long-term monetization
• Not every registration paid for itself.
• Adjust reported roughly 10% more registrations than TikTok Ads Manager.
The biggest lesson?
Tier-3 isn't just cheap traffic.
It's one of the best places to test creatives, collect data, and find winners before scaling elsewhere.
Great performance is usually the result of a lot of testing not luck.