Claim and evidence
Switching costs shift accordingly.a16z argument · a16z.com · 2026-05-14
Anecdotal
Practitioner experience without published numbers.
Working · 2026-05 Forward-looking argument from May 2026 essay, no retention data.
What the numbers show
No published number for this tactic has passed our word-for-word check yet, so there is no chart.
Why it works
If switching data is cheap, stickiness moves to the learned context that makes the assistant useful.
How to run it
- Store user preferences and decisions
- Let the AI act on that context
- Show value of accumulated context
- Offer export without breaking the experience
Ways to run it
Design options to test, not promised results.
- Option AStored preferences
Save the user's preferences and past decisions and apply them automatically.
- Option BVisible accumulated context
Show users what the assistant has learned about their work so the value is easy to see.
- Option CExport without breaking
Let users export raw data freely while the learned context keeps the experience useful.
Dated examples
- 2026
a16z argument
Essay predicts switching costs shift toward the AI orchestration layer.
“Switching costs shift accordingly.”
✓ quote checked word-for-word on the pagea16z.com · 2026-05-14
Does it still work?
Forward-looking argument from May 2026 essay, no retention data. source
Risks
- Speculative
- Privacy and portability expectations
Sources
- a16z.com
“Switching costs shift accordingly.”
✓ quote checked word-for-word on the page
Change log
- 2026-11-20 — harvest v0.1 (retention)