Honestly, it depends on two things, and one of them is you.
The first is traffic. Personalisation works on patterns, and patterns need volume. If your site sees a few dozen visitors a week you will not be able to separate a real improvement from an ordinary quiet fortnight. A few hundred a week is where the numbers start to mean something.
The second is content. Spectare chooses between the things you have written. If everything you have says the same thing in the same way, choosing between them changes nothing. The gain comes from having genuinely different answers for genuinely different visitors: someone comparing you against a competitor needs a different paragraph from someone who has never heard of you.
What tends to move first is not the headline conversion rate. It is the drop-off on the page where people used to leave, because the section they landed on finally spoke to the reason they came.
If you want a number before you commit, the honest answer is that nobody can give you one, because it depends on content you have not written yet. What you can do is take the page that already gets the most traffic, put one varying section on it, and read the difference after a month.
This is one atom from the Spectare content library. Spectare assembles the right atoms for each visitor in real time, based on who they are and how they arrived.