I think the upwards pressure from companies not being able to hire enough good engineers is likely a much stronger force. That being said, we're not tied to the % model. We plan to experiment with both our selection process and pricing to figure out what works best.
One cool innovation I've seen is converting from a % upfront model to a % recurring model (Ex. instead of 25% upfront convert to 2% over 12 months).
It benefits start-ups because they don't have a large cash outlay at once and aligns the incentives of the firm to find great candidates that will stick. It also creates a more predictable revenue stream in the form of monthly recurring revenue.