A monthly agreement where a client prepays for a set number of your hours. The client gets guaranteed capacity and priority; you get predictable recurring revenue. Retainers are priced below your ad-hoc rate (typically 10–20% off) in exchange for the commitment — the discount is the price of predictability, not a sale.

The retainer also changes the relationship dynamic: you stop being a vendor they call in emergencies and become infrastructure they plan around. That is why retainer clients churn less, refer more, and accept rate increases more gracefully — they are buying continuity, and continuity is the hardest thing for a competitor to displace.

10–20% is the healthy band. Below 10% the client sees little reason to commit; above 20–25% you are repricing yourself, not rewarding commitment, and every future negotiation anchors to the lower number. The calculator flags discounts above 25% as too deep. Sell the guarantee (priority access, reserved capacity, faster turnaround) — the discount is the sweetener, not the product.

Prefer use-it-or-lose-it; cap any rollover. Full rollover turns unused hours into deferred work you owe — a quiet month becomes a 2× workload month at half the effective rate. If a client insists, allow partial rollover (50%) for one month only, tracked in writing. The policy must be in the agreement before the first invoice, not negotiated in the third month.

There is a softer middle ground some freelancers use: "rollover with expiry" — unused hours roll for 30 days, then lapse. It feels generous to the client while mathematically converging to use-it-or-lose-it for anyone who is not chronically under-using. Whatever you choose, the hours ledger should be on every invoice: "20 hrs reserved, 14 used, 6 lapsed" — transparency prevents the end-of-quarter argument.

Overage bills at your full ad-hoc rate — never at the retainer discount. State this in the agreement: "hours beyond the retainer are billed at $X/hr". Without an overage clause, scope creep eats the retainer from the inside — the client gets 30 hours for the price of 20 and your effective rate collapses. (See the scope-creep cost calculator for the math.)

Enough to cover your baseline costs, few enough to stay flexible. A common structure: retainers cover 40–60% of your capacity (your "salary"), ad-hoc projects fill the rest at full rate (your "profit"). Selling 100% of your hours on retainer caps your income at the discounted rate — leave headroom for full-rate work and growth.

Lead with their savings and your availability, not the discount. "You used 18–22 hours a month for the last quarter at $100/hr — about $2,000. A 20-hour retainer is $1,700/month: you save $300+, and your work jumps the queue." The calculator computes both numbers for exactly this conversation. Add the overage and rollover terms in one page — simplicity closes.

Monthly for flexibility, quarterly for commitment. Monthly retainers are easier to sell and easier to exit; quarterly retainers (often with a slightly deeper discount or a paid-upfront term) give you a longer planning horizon. Avoid annual lock-ins at discounted rates — your rate should rise yearly, and a 12-month discount freezes it.

They lower it by the discount — deliberately. A $100/hr rate at 15% off is an $85 effective rate on retainer hours. That is fine as long as it is a choice: the discount buys revenue stability. It becomes a problem only when retainers crowd out full-rate work — monitor the blend monthly (see the effective hourly rate calculator).

Put a pause clause in the agreement upfront. Typical terms: one pause per year, up to 2 months, with 30 days' notice — and paused months don't bank hours. Without a clause, "pausing" becomes an open-ended free option on your capacity. Notice periods protect your pipeline: 30 days lets you backfill the slot.

Yes — reframe the retainer as reserved capacity, not hours. A designer can sell "priority design days": 4 reserved days/month at a set price. A writer can sell "2 articles/month, delivered by the 15th". The unit changes; the economics don't: guaranteed volume at a modest discount in exchange for guaranteed revenue. Define the deliverable unit crisply or scope arguments follow.

Annual review, 30–60 days' notice, in writing. "Effective January, the retainer moves from $1,700 to $1,850/month reflecting current rates — still 15% below ad-hoc." Clients accept scheduled increases far better than surprise ones; build the review into the original agreement ("rates reviewed annually") so it is expected, not sprung.

Anchor the increase to something real: your rising ad-hoc rate, expanded scope, or simply the annual review clause they signed. And re-run this calculator at the new rate — showing the client that the discount percentage is unchanged ("still 15% below my current $110/hr") reframes the increase as consistency, not a price hike. Clients leave over surprises, not over scheduled 5–8% annual adjustments.

Both, in layers. Retainers are your salary — baseline revenue that covers costs. Projects are your profit — full-rate work that grows the business. The healthiest freelance practices run 40–60% retainer, 40–60% project. Pure retainer caps income at a discount; pure project means feast-or-famine. The calculator shows what one retainer contributes — multiply by your retainer count for the full picture.