MRR (monthly recurring revenue) and ARR (annual recurring revenue) are popular buzzwords for good reason.

Recurring revenue can change a business. It helps solve some of the biggest problems agency owners deal with:

• Financial stability. You know what you’ll make next month without having to sell again.

• Reliability. You can plan longer-term hires, tools, and cash flow instead of reacting.

• Valuation. Predictable revenue makes an agency easier to sell, or easier for you to keep running

When I sold my agency SuperFriendly, we didn’t have any recurring revenue. That was partly because the agency was already shut down, but also because that wasn’t how we ran it. Looking back, I could have sold it for much more if it already had recurring revenue in place.

So it makes sense that agencies chase retainers, one of the most common forms of MRR.

The problem isn’t retainers themselves. The problem is how most retainers are sold.

The most common retainer trap

Most agencies sell retainers as blocks of time.

The moment you sell time, you create a problem. Time by itself isn’t valuable. We all have the same amount of it. You can’t buy more of it. You can’t really sell it. What matters is what you do with that time.

So you explain what you’ll do with the time. And what does that usually become?

Maintenance.

“We’ll do maintenance for X hours per month.”

But let’s be honest. Most modern websites don’t need 10–20 hours of real maintenance every month. Not forever. Not consistently. Not really.

What happens next is predictable:

• You invent work

• You track and report hours

• You defend invoices

• You manage expectations

• You argue about edge cases

All to protecting yourself from the consequences of selling time. (It’s also why I think calling yourself “fractional” does more harm than good.)

Here’s the key pattern most people miss:

If your retainer requires complicated rules, you already lost.

Rollovers. Caps. Overages. Ranges. Points. “Use it or lose it.”

These aren’t smart pricing systems. They’re guardrails built around a flawed unit of sale.

A better retainer model: selling insurance

There’s a simpler and more honest way to think about retainers:

As insurance.

People don’t buy insurance because they expect to use it every month. They buy it because not having it is too risky.

Insurance is built to be something you ideally never need, except in rare or serious cases.

That’s not a weakness of the model. That’s the entire point.

Insurance math

Insurance companies do handle claims. That includes real work, vendors, and labor.

What they don’t promise is constant or guaranteed work for every customer.

They sell coverage, not capacity.

They can do that because demand is predictable across many customers, even if it’s unpredictable for one customer.

Insurance companies rely on:

• Actuarial tables

• Historical loss data

• Probability distributions

• Risk pooling across millions of customers

They don’t guess. They don’t renegotiate or reassess every month. They don’t worry about whether one customer “used what they paid for.”

They know—with high confidence—roughly how many claims will happen, how big they’ll be, and how much work they’ll require overall.

The system is designed around probability, not fairness per customer.

Compare that with typical agencies. Agencies sell retainers as if:

• Every client is entitled to full utilization

• Every month

• Whenever they want

Which would put the operation out of business almost instantly.

Like insurance companies, agencies are already running a risk pool. But unlike insurance companies, they refuse to realize, acknowledge, and/or admit it.

Instead of planning across all clients, they promise time to each one. Instead of pricing for probability, they price for entitlement. Then they spend the rest of the relationship apologizing.

Insurance companies don’t apologize when nothing happens. Agencies do.

That’s the difference.

Fairness in the wrong place

Look at how retainers are usually explained. Everything is about being fair:

• Fair if the client doesn’t use all their hours

• Fair if the work takes less time than expected

• Fair if nothing urgent comes up

Insurance doesn’t work like that.

You don’t get a refund because you didn’t crash your car. You don’t get rollover claims because it was a quiet year. You don’t get bonus coverage because the insurer had spare capacity.

Retainers fail when fairness is measured in usage instead of risk.

Not using your insurance is a sign that things are going well. Can your client say the same of not using their retainer?

What great retainers actually sell

Great retainers aren’t about selling tasks. Or hours. Or deliverables.

Great retainers sell three things:

1. Priority. When something matters, you’re first.

2. Preparedness. Someone already understands your system, context, and risks.

3. Reduced downside. Fewer fires, faster recovery, less exposure when things break.

None of those map cleanly to hours.

Availability vs. Coverage

Another common retainer pattern: agencies tend to talk about availability instead of defining coverage.

That’s why clients assume:

• You’re always “on”

• You’ll squeeze things in

• You work on their schedule

Insurance doesn’t promise availability. It defines coverage.

Clear triggers. Clear limits. Clear exclusions.

No negotiation every month.

When retainers feel heavy, it’s usually because availability was vague and coverage was never defined.

Reframing retainers

Stop calling it maintenance. Stop selling time. Stop building elaborate rule systems to defend yourself.

Start selling protection.

Protection from what?

Business risk.

Lost revenue. Lost leads. Lost credibility.

Like insurance policies, the best retainers are the ones people pay for and hope they never need.

Which is exactly why they keep paying.

How to pitch insurance

Here’s a starter script you can use to pitch insurance to your clients:

Hey { CLIENT NAME },

With your site launch a few months away, I was thinking about how we might reduce risk for you going forward without adding more process or overhead.

What usually causes the most damage in high-stakes moments like launches or exec reviews isn’t the issue itself; it’s the delay. A platform change or third-party update exposes something no one could have predicted, and fixing it turns into a round of approvals, scope questions, and quotes.

We started offering protection plans last year to remove that friction. If traffic drops, a deploy goes sideways, or an external change creates urgency at the worst possible time, you’re not pausing to figure out next steps. You have clear coverage and priority when it matters.

Most months, you won’t need anything at all, because it’ll be our job to make sure of it.

It’d only be about 10% more than your current plan with us, and we could add it on immediately.

Can I show you how it’d work on our call on next Wednesday

Hey { CLIENT NAME },

With your site launch a few months away, I was thinking about how we might reduce risk for you going forward without adding more process or overhead.

What usually causes the most damage in high-stakes moments like launches or exec reviews isn’t the issue itself; it’s the delay. A platform change or third-party update exposes something no one could have predicted, and fixing it turns into a round of approvals, scope questions, and quotes.

We started offering protection plans last year to remove that friction. If traffic drops, a deploy goes sideways, or an external change creates urgency at the worst possible time, you’re not pausing to figure out next steps. You have clear coverage and priority when it matters.

Most months, you won’t need anything at all, because it’ll be our job to make sure of it.

It’d only be about 10% more than your current plan with us, and we could add it on immediately.

Can I show you how it’d work on our call on next Wednesday

Hey { CLIENT NAME },

With your site launch a few months away, I was thinking about how we might reduce risk for you going forward without adding more process or overhead.

What usually causes the most damage in high-stakes moments like launches or exec reviews isn’t the issue itself; it’s the delay. A platform change or third-party update exposes something no one could have predicted, and fixing it turns into a round of approvals, scope questions, and quotes.

We started offering protection plans last year to remove that friction. If traffic drops, a deploy goes sideways, or an external change creates urgency at the worst possible time, you’re not pausing to figure out next steps. You have clear coverage and priority when it matters.

Most months, you won’t need anything at all, because it’ll be our job to make sure of it.

It’d only be about 10% more than your current plan with us, and we could add it on immediately.

Can I show you how it’d work on our call on next Wednesday

No hourly math. No maintenance theater. No “on-call” guarantees.

Just insurance.

For your client and your business.