Four things we won't do.
Saying yes to a feature is easy. Saying no defines a product. Here are four positions we hold when the temptation to ship anyway shows up.
1. We won't host your client data
Trainerize, TrueCoach, HubFit, and Everfit all store your client history on their servers. Your access ends when your subscription ends. Coaches we interviewed mentioned this almost always with the same word: hostage.
Coach Sheet writes every client file to a folder in your own Google Drive. The folder layout
mirrors what working trainers already use: a master template called # Control Center
(the # sorts it first), then numbered per-client folders. If you cancel us, the
folder structure stays intact, the formulas keep working, you keep using Sheets normally. The
only thing that goes away is the menu bar we install. That's the inverse of "hostage" by design.
2. We won't charge per client
Per-client pricing punishes growth. A trainer at 50 clients on Trainerize pays roughly $90 a month for Studio. At 100 clients, $180. The work the software does is functionally identical at either roster. The bill grows because the software vendor extracted leverage from your career milestone.
Coach Sheet is $15. Whether you're at 10 clients, 50, or 100. The math runs: at 50 clients, a coach saves about $900 a year switching to us from Trainerize Studio. We can do this because Apps Script execution costs are negligible and we don't need to fund a sales team to keep you in step-up tiers. It's a deliberate margin choice, not a loss leader. See the full pricing breakdown.
3. We won't ship "AI workout generator"
Every coaching SaaS in 2026 has a "generate workout with AI" button. We surveyed 138 Reddit threads asking what trainers actually use. Zero of those threads asked for it. Many of them mentioned it disparagingly: "I don't trust AI to program for my client" was a recurring sentiment.
The job of a coaching tool is to remove minutes from data entry, not minutes from clinical judgement. If you want a tool that drafts mesocycles from a prompt, that exists. We're not a competitor in that lane. We are a competitor in the lane where the trainer is the expert and the tool is the scribe. Why we hold this line.
Caveat. We use formulas everywhere: Mifflin-St Jeor for BMR, Epley/Brzycki for e1RM, three-day moving average for weight smoothing. Math is not AI. Showing a number that follows deterministically from the inputs is not "magic". We are firm on the line: deterministic math, yes; generative model picking exercises, no.
4. We won't hide formulas
Many tools display "TDEE: 2,264" with no indication of how that number was reached. Coach Sheet displays "TDEE: 2,264 (BMR 1,794 × Activity Multiplier 1.55, Mifflin-St Jeor)" with a hover tooltip on the formula. This is not a UX flourish. It's an editorial position.
You should be able to inspect every calculated number and trace it back to its inputs. RPE colour bands? Configurable: yellow can mean RPE 7-8 or 8-9 depending on the coach. Volume landmarks (MEV/MAV/MRV)? Configurable per muscle group. Macro split ratios? Configurable (40/30/30 vs 30/30/40 vs your custom). The defaults follow Helms / Schoenfeld / Israetel conventions because those are the most cited; the editing happens in the cell.
Things we'll change our mind about
These positions are firm but not religious. We update them publicly when evidence accumulates.
- Native iOS / Android apps
Re-evaluate when 100+ paying coaches request it. PWA covers 80% of native UX with zero app-store review delay.
- Whitelabel / branded apps
That's a $300/month feature, not a $15/month feature. We won't bundle it into the base price.
- English-only content
We will translate the cornerstone surface to Ukrainian first (the team's home market), then Polish where SEO data is encouraging. Not via machine translation alone.
- Open source the Apps Script
Likely yes, eventually, with a non-commercial clause. Trainers should be able to read what runs in their Drive. We're not there in week one.