The three models, and what each one rewards
| Model | You pay | It rewards the vendor for |
|---|---|---|
| Percentage of ad spend | A share of what you spend, typically monthly | Spending more of your money |
| Per seat | A fee per user with access | Adding users, not results |
| Banded subscription | A flat fee set by the size of the account | Keeping you past renewal |
None of these is dishonest. They just point in different directions, and the direction matters more than the headline figure, because you will be inside one of them for years.
Percentage of spend deserves the hardest look. It is the model most agencies use and several platforms have inherited, and it has a structural problem: the recommendation to increase budget comes from the party paid more when you do. Good operators manage that conflict honestly. It is still a conflict, and it is worst exactly when you are scaling, which is when advice matters most.
Per seat pricing has a quieter failure. It prices access rather than value, so the rational response is to limit who can see the system, which is the opposite of what you want from something making changes to your accounts.
What actually drives the price
Whatever the model, the cost tracks one of three things:
- The size of the account. More spend and more traffic mean more decisions, more data and more consequence when something is wrong.
- The number of channels. Each connected platform is its own integration, its own API and its own failure modes.
- What the system is allowed to do. Some vendors charge more for the autonomous tier and sell the read-only version cheaply.
That third one is worth checking before you sign anything. If automation is the upsell, the base product is a dashboard, and you will be paying twice to get the thing you came for.
How So What Labs prices it
Three tiers, set by the size of the account rather than by capability. You land in a tier on whichever of monthly visits or monthly ad spend crosses the threshold first, because a B2B advertiser can run modest traffic against heavy spend and would otherwise be under-tiered against the work involved.
Every tier runs the full system. Autonomy is not behind the paywall, and there is no version that watches without acting, because that product is a report and you already have several. Apps can be bought individually or as a suite, where the suite costs less than the same apps separately. Annual billing is ten months paid for twelve. Enterprise is quoted, because multi-brand and agency accounts vary too much to band honestly.
The figures are on the pricing page and are deliberately not repeated here, so that this answer cannot go stale while still sounding confident.
The costs that are not on anyone's price list
- Fixing the tracking. Most accounts need work here before automation is safe, because every automated decision inherits the measurement error underneath it. Budget for it.
- The first month. Simulation, disagreement, tuning the guardrails. Real time from a real person.
- The data work. Getting margin or lifetime value into the system is what makes the whole thing worth more than bid management, and it is rarely a switch.
- Whatever you keep paying for. Platforms replace some tools and not others. The saving is only real if something actually gets cancelled.
Questions worth asking a vendor
- Is the price tied to my ad spend, and if so what happens when I double it?
- Is autonomy included at every tier, or is it the upgrade?
- What is the minimum commitment, and what does month one look like?
- Is implementation included?
- What does the price do when I add a channel?
Cheaper than the alternative is a different question, and it depends on what you are comparing against: an autonomous platform against an agency retainer is not a straight fee comparison.