How to negotiate a reserve with Stripe (the real playbook)

Stripe reserves are not a negotiation in the traditional sense. There is no sales rep with discretion. There is a risk model, a risk-ops team that reads its output, and an underwriting team above them. Your job as the operator is to move the inputs the model uses — not to argue with the output.

This is the script we use when one of our portfolio brands hits a reserve on Stripe, and the sequence that consistently compresses the hold from 25% to 10% (or occasionally zero) within 60 days.

1. Understand what triggered the reserve

Stripe almost never volunteers the specific trigger. You will get a boilerplate message about "risk". The real triggers are narrow and identifiable:

• Chargeback ratio trend — not the snapshot, the trend. A ratio climbing from 0.4% to 0.7% over 30 days triggers a reserve even if you never hit threshold.
• Refund velocity spike — refunds over 8% of gross in a 7-day window reads as product issues or friendly-fraud suppression.
• Subscription churn cliff — cancellation rate tripling in a week reads as a deliverability problem.
• Descriptor complaints — descriptor inquiries that turn into disputes spike the fraud score.
• Vertical reclassification — Stripe quietly re-scored your MCC into a higher-risk bucket after a website crawl.

Figure out which one is yours before you open a case. We dump 90 days of dispute, refund, and cancellation data into a spreadsheet and look for the inflection point. The inflection point is almost always the trigger.

2. Pull the four levers the model actually reads

Stripe's reserve engine weighs about a dozen signals, but four of them move fast enough to matter over a 30-60 day negotiation window:

Lever one — chargeback trend reversal

The model looks at 30/60/90-day chargeback ratio. A declining trend is the single strongest input. Operators who cut their ratio from 0.8% to 0.35% over 45 days routinely get reserve reductions at the next review. See our chargeback ratio guide for the specific tactics.

Lever two — dispute-win rate

Winning disputes with compelling evidence changes how Stripe reads your chargebacks. A 55% win rate is the unofficial threshold where the model starts treating disputes as recoverable rather than pure loss.

Lever three — subscription mix and dunning

Subscription-heavy accounts get better reserve treatment than one-shot accounts at equal GMV, because recurring revenue is less fraud-prone. If you're running one-shot + subscription, increase the subscription share and the reserve math improves.

Lever four — descriptor hygiene

Move to a dynamic descriptor that includes brand + support phone. Descriptor complaint rate is a fraud-score input that operators routinely ignore.

3. Run the escalation path in the right order

Stripe's support structure has three tiers and they do not talk to each other unless forced.

Tier one — Radar review request

First action: open a Radar fraud-score review in the dashboard. Provide your own fraud analysis. This lands with the risk-ops triage team and creates a documented case number. Most operators skip this and go straight to chat. That is wasted time.

Tier two — risk-ops case

After 72 hours with no Radar response, open a risk case through the dashboard "Contact risk" flow. Reference your Radar case number. Attach your chargeback data, dispute-win rate, refund trend, and a one-page remediation plan. The remediation plan is load-bearing — risk-ops reviewers look for operator self-awareness.

Tier three — underwriting escalation

If risk-ops declines to reduce the reserve, request underwriting escalation. Underwriting reviews the whole account structure, not just the specific trigger. This is where operators with legitimate multi-brand businesses sometimes get reclassified into a lower-risk profile.

The thing that does not work: emailing the account manager. Stripe account managers above certain volume thresholds exist but they have almost no discretion on reserves. They'll route you back to risk-ops anyway.

4. The 30-60-90 remediation plan template

The remediation plan you attach to your risk case is a short document. It should contain exactly these sections:

• Baseline metrics (past 90 days) — chargeback ratio, dispute-win rate, refund rate, cancellation rate, trailing 30-day GMV. Be honest. Risk-ops cross-checks against their own data.
• Root cause — one paragraph on what caused the metrics to move. If you don't know, say so — pretending is worse than admitting.
• 30-day actions — specific changes (e.g., "3DS challenge flow added for non-US cards, descriptor updated to brand+phone").
• 60-day targets — metric targets you commit to (e.g., "ratio below 0.5%, win rate above 55%").
• 90-day review trigger — request for reserve review on a specific date.

This document gets attached to the case. It also becomes the evidence you reference at the 90-day review. Operators who write it treat themselves like a real business. Operators who don't treat themselves like a support ticket.

5. What compression looks like on the other side

Typical trajectory for an operator who actually executes:

• Day 0 — 25% rolling reserve, 90-day hold
• Day 30 — first review, usually no change (risk-ops wants 60+ days of data)
• Day 60 — reserve reduced to 15% if metrics are holding
• Day 90 — reserve reduced to 10% or removed, hold period shortened to 60 days
• Day 180 — if clean, reserve fully removed, standard payout schedule restored

If none of that happens by day 90, the reserve is structural — meaning Stripe's model has classified you as a permanent elevated-risk account. At that point the question stops being "how do I get Stripe to reduce my reserve" and becomes "is Stripe the right home for this business." See our when to fire your processor guide.

6. Multi-brand operators — different math

If you're running multiple brands on separate Stripe accounts, each account has its own reserve negotiation. That's N fights, N reviews, N remediation plans. It is also N chances for a single brand to tank the reserve for the portfolio because Stripe's risk model does cross-reference related accounts.

A parent merchant account with sub-brand descriptors consolidates this into one underwriting relationship. Fewer reserve negotiations, one aggregated chargeback picture, shared dispute-win evidence. That's the structural tradeoff we cover in multiflow vs Stripe and the true cost of 15 Stripe accounts.

What not to do

• Don't open three support chats on the same issue — they merge your cases and reset priority.
• Don't threaten to leave — Stripe has heard it from every operator. It changes nothing and sometimes worsens the reserve.
• Don't argue the trigger — the model sees what it sees. Focus on changing the inputs.
• Don't pay an "account rescue consultant." Almost all of them send the same templated email you could send yourself.

What to do next

Pull your 90-day metrics today. Write the one-page remediation plan. Open the Radar case. Then wait 72 hours before the next move. Stripe reserve timelines are structural — acting faster than the system can read your data doesn't help you.

If you operate more than three brands, the reserve problem compounds across accounts. Our 12-question application helps us see whether the parent-account model fits your situation.