Retainers: sell a monthly product, not discounted hours
Notes from 2017, when I was working out what a Siftware retainer should be. Much of the thinking came from Brennan Dunn's guide to retainer agreements; the opinions are mine.
The trap
"Ten hours a month, paid in advance, at a small discount." It sounds like a retainer.
The client works out what the hours cost, decides they'll just pay you ad hoc instead, and you're back where you started.
If you do sell hours, roll-over makes it worse. Clients count the hours obsessively, or go quiet for months and then drop a pile of banked-up work on you at the discounted rate.
Four kinds
- Time: hours or days a month. Fine for fiddly support, or the requests that come in after a project has ended. Hard to sell, for the reason above.
- Project: the project's cost spread over its expected length and charged monthly. The client can budget, and they're paying for deliverables, not your time.
- Value: the problem is clear, the solution isn't yet; a monthly fee for steadily reducing a costly problem. Works best when you're retained exclusively for a period. Very hard to sell unless you're seen as the expert.
- Productised: a range of services packaged as a monthly product. This is the one.
What goes in a productised retainer
A retainer lets you improve, insure and train (Brennan says optimise, insure and train). In practice:
- Some support, a set number of requests a month that don't roll over.
- Housekeeping and maintenance you do without being asked: updates, checking the backups still work.
- A monthly report on numbers the client cares about, ideally ones you can put a value on. It also shows you care about their business month after month.
- Access to tools you get cheaper through scale.
- Managing other suppliers for them. Taking direct reports off a busy manager's plate is worth a lot.
- First call on your time when you schedule.
Leave big blocks of effort out of the standard package and sell them as add-ons. They don't scale without a team, and they make the retainer look expensive.
Selling it
- Tiers. A basic one, a middle one, and a premium one anchored against hiring the role in-house. The client chooses between A, B and C instead of yes or no.
- Make it easy to cancel. It lowers the barrier to signing, and if they're splitting hairs over the monthly deliverables you probably don't want them anyway.
- A simple contract: what they get each month, the cost, when they pay, how long it runs, how soon they can cancel, what happens in between, communication and working hours.
- When: once they're already a client. Mention it in the proposal, then bring it up again as the project wraps up.
This is the other half of maintenance is a market you have to make.