# CacheFold Sales Sheet

## Test whether one Redis/Valkey namespace can support a $50k/year savings case.

CacheFold is for teams with large Redis or Valkey memory pressure in deterministic,
repeated data: catalogs, config, pricing, localization, rules, or similar reference
data. It assesses whether one namespace can reduce memory pressure, delay capacity
expansion, and justify a controlled pilot.

The first step is a free async fit screen. If the screen looks credible, the paid
assessment is scoped to one namespace at $7.5k-$10k.

## What CacheFold Assesses

- In-scope baseline GiB, replica/headroom factor, and Redis/Valkey pricing basis
- Top prefixes by memory and whether values are deterministic and repeated
- Expected net Redis memory reduction after parameters and exceptions are counted
- p99 budget, read/write behavior, TTL/churn, and source-of-truth path
- Validation owner, rollback owner, and pilot acceptance criteria

## Buyer Outcome

The assessment produces a PASS/FAIL recommendation with:

- quantified savings range based on buyer-provided infrastructure inputs
- measured parity plan and p99 budget
- source-of-truth and rollback requirements
- recommendation on no-bid, diagnostic, assessment, or pilot path

The savings case is buyer-specific. CacheFold should not be treated as a generic
Redis tuning service or as a fit for sessions, tokens, per-user state, high-entropy
blobs, or high-churn namespaces without a clear validation path.

## Proof Posture

CacheFold is built around artifact checks:

- baseline comparison against the original Redis/Valkey behavior
- reconstruction audit with zero-mismatch requirement
- mixed read/write checks, including stale-after-write and restart consistency
- p99 measurement against the agreed budget
- rollback path before pilot

The public proof posture is confident about the artifact path and careful about
buyer-specific savings. The assessment determines whether the namespace, latency
budget, and economics support a pilot.

## Engagement Path

1. Free async fit screen: platform, top prefixes/GiB, value shape, read/write
   ratio, TTL/churn, source-of-truth check, p99 budget, validation owner, rollback owner.
2. Paid assessment: $7.5k-$10k for one namespace, ending in PASS/FAIL.
3. Pilot after PASS: $15k-$25k for shadow, canary, gated rollout, and rollback drill.
4. Annual license: max($30k, 10-15% of validated first-year net savings).

## Fast Fit Packet

Send:

1. Redis/Valkey platform and deployment shape
2. Rough in-scope memory footprint in GiB, including replicas or headroom
3. Top prefixes by memory
4. Value shape and source-of-truth path
5. Read/write mix, TTL/churn, and p99 budget
6. Validation owner and rollback owner

If the data and economics look weak, the recommendation should be no-bid or a
small diagnostic rather than a paid assessment.