Customer.io email marketing agency: what a month of retainer should actually ship
In December 2009, Domino's Pizza did something most companies avoid: it admitted, on television, that it had heard the complaints. Consumers had kept saying the pizza "tasted like cardboard". After about 18 months of development and 36 combinations of crust, cheese and sauce, a new recipe rolled out to every store on 27 December 2009. The new chief executive, Patrick Doyle, went on camera to admit he had heard what the focus groups said (Nation's Restaurant News).
On 4 May 2010, Domino's reported domestic same-store sales up 14.3% for the first quarter. Doyle said: "A remarkably better pizza, and our honesty in how we told consumers about it, is paying off" (Domino's results release). Asked how much of the jump the new pizza accounted for, Doyle replied: "All of it" (Nation's Restaurant News).
Most agency retainers never have their focus-group moment. The monthly report shows opens, clicks and a list of tasks, and nobody says the plain thing: nothing live is moving the number the client cares about. This post sets out what a month of a Customer.io email marketing agency retainer should ship, and the test that tells you by month three whether yours is.
TL;DR: A Customer.io email marketing agency retainer should put something live every month that you can name: a new automation, or a named change to one. Each piece is tied to one metric, with a baseline taken before it launches. The monthly report shows that metric against its baseline, ahead of opens, clicks and a task list. Set each automation's goal in Customer.io before it goes live, because the docs say goals are not applied to people who entered earlier. If month three ends with no live automation moving your priority metric, the retainer is buying meetings, not results. Clients agree: in Setup's survey, 48% of clients named dissatisfaction with delivery as their top reason for leaving an agency, and agencies ranked it seventh. NerveCentral retainers run from £2,000/$2,900 to £12,000/$17,500 a month, depending on requirements.
What a month should ship
A retainer is a monthly fee for ongoing work. Clutch describes it as a fee paid with the intention of a long-term relationship across "multiple projects and campaigns" (Clutch). Customer.io describes partner work the same way: its partner team helps agencies "grow service revenue with implementation and retainer services" (Customer.io partners).
Neither definition says what the money should produce. Here is our answer. A month of a Customer.io retainer should ship three things:
- Something live you can name. A new automation in your workspace, sending to real people. Or a named change to a live one: a new email added, a branch rebuilt, an exit fixed. We worked on onboarding does not count.
- One metric for each piece. Every automation names the step in the customer's journey that it moves and the window it is measured in. An abandoned checkout flow is measured on checkouts recovered, not on opens.
- A report against a baseline. The number before the automation went live, the number now, and the dates of both.
Everything else a retainer pays for (calls, strategy, copy reviews, deliverability checks) exists to produce those three. It matters, but it is not the output.
Why retainers drift into meetings
Retainers drift because most are priced and scoped by effort, not output. Ignition surveyed 273 US agency leaders in April 2025. 28% said their main pricing model was hourly billing, and another 28% said subscription or package pricing (MarketingProfs on Ignition's survey). Ignition's own summary adds that 57% of agencies lose $1K to $5K a month to scope creep (Ignition).
Neither model is wrong, but both measure what goes in: hours worked or emails written. Neither tells you whether anything moved.
Hours go less far than buyers expect. Clutch puts the average email marketing agency rate at $100 to $149 an hour (Clutch). By our arithmetic, $2,900 a month at those rates buys roughly 19 to 29 hours. Calls and reports can use much of that.
Clients notice. Setup's sixth Marketing Relationship Survey, released late 2024, drew over 400 contributions from brands and agencies. Clients ranked dissatisfaction with delivery as the top reason they end agency relationships, cited by 48% (Setup). Agencies put it seventh on their own list. They blamed client leadership changes and budget cuts first. The same survey found that 40% of clients expected to switch agency partners within six months.
Agencies blame forces outside their control. Clients say the work did not arrive.
One metric, set before the first build
The fix starts before anything is built. Every NerveCentral client has one priority metric, written down with four facts: the baseline, the date the baseline was taken, the target and the target date. If those are not agreed yet, the plan says so and names the piece of work that agrees them. Nothing hides behind we will figure out measurement later.
One metric sounds narrow. That is the point: a retainer reporting twelve metrics can always find one that went up. Our lifecycle marketing scorecard makes the same case against reporting on open rates.
Customer.io gives you the tool to hold each automation to its metric: the goal. A goal is "what you want your audience to accomplish during your automation", defined by conversion criteria (Customer.io docs). Three details decide how a retainer should use it.
- Set the goal before launch. If you add conversion criteria to a live automation, they apply only to people who enter it after you set the goal. Customer.io will not count conversions retroactively. An agency that builds first and adds goals later has lost the first weeks of evidence.
- The window is capped. You can count conversions for up to 90 days after a delivery. Pick a window that matches the metric, not the longest one available.
- Automations do not share credit. Customer.io counts conversions independently for each automation. A person who goes through two automations with the same goal converts in both. Add up the automation conversions in a report and you will overstate what email did. Our post on the goals and conversion events double-count trap covers this in detail.
So the monthly report should never sum automation conversions into one headline. It should show the priority metric from the business's own numbers, against its baseline, with each automation's contribution beside it.
What the first 90 days should look like
A retainer at NerveCentral can open with a 90-day plan, signed off by the client before any build starts. After it, the work is planned month by month from the monthly review. The plan has a fixed shape.
The map. Every automation type in our campaign catalogue appears on the map, grouped by where it sits in the customer's journey: before they pay, when they pay, after they pay, and regular sends. Each row is marked Build (with the month it goes live), In place, After 90 days, or Not needed with the reason. Nothing is left off, so the client sees the whole picture and the next 90 days start from it.
The order. Build order goes by value. Automations that touch revenue come first, billing protection second, retention third. In practice, abandoned checkout and the automations that turn sign-ups and trials into first payments usually land in month one or two. Sending setup and domain warm-up is always the first line of month one, because nothing else matters if the email does not arrive.
The capacity. Each month stays inside the capacity in the signed proposal: a set number of emails and automations. That number is what the client is paying for, so it is written down.
Here is an illustrative 90-day plan for a subscription business. The automations and their order change with the client's data and journey.
| Month | What goes live | Measured as |
|---|---|---|
| 1 | Sending setup, then abandoned checkout and trial ending | Checkouts recovered; trial-to-paid rate |
| 2 | First purchase offer and failed payment | First payments; charges recovered within the retry window |
| 3 | Usage drop and win-back | Customers active again; customers reactivated |
After sending setup, each item is a named automation with one metric and a baseline taken before it launched.
If your workspace is already full of automations, the map starts with what is there. A Customer.io audit finds which live automations are working, which are broken, and which have exit conditions that never fire. Fixing a live automation is often the fastest win of month one.
How one live sequence kept earning: Slimfy
The clearest example of what a retainer month should look like comes from our work for Slimfy, a supplements brand. They had traffic and no email marketing (case study).
Cart abandonment came first, because, as the case study puts it, it is usually "your low-hanging fruit". After researching the product, the competitors and the audience, we wrote a 6-part sequence. Opens and clicks were watched, but sales were the metric that counted: each email was judged on sales conversion rate and earnings per click. After three months, the sequence was averaging an additional $27,000 per month in revenue. The client says the revenue is over 10x their investment.
We also kept adding emails to that one sequence, and it reached 9 emails before people stopped clicking through. That depth was found one email at a time, not planned on paper.
That is what a good month after the build looks like: a named change to a live automation, measured against the number it already produces. A month that adds email seven and shows what it earned counts as shipping. A month that reviews the sequence does not.
What the client should see every week and month
Output needs a rhythm, or it slips. Every NerveCentral client gets the same one, whether the engagement is a build or a retainer.
| When | What the client gets |
|---|---|
| Every Friday | A status update: what shipped, what the numbers say, what comes next |
| Every week (or fortnight, where agreed) | A call |
| First week of the month | The monthly report |
| After the report | Next month's work, each piece naming the metric it moves |
The monthly report always has the same shape, so one month compares with the next. The priority metric comes first, against its baseline and target. Then each automation with its numbers. Then what we did, what we do next month, and what we need from the client.
Inside NerveCentral, the report is followed by a keep, stop, start review. We list every automation running for the client and what it contributed to the priority metric. Each one gets a decision: keep it, stop it, or change it. We propose up to three new starts, and each must name the metric it moves. Anything running that cannot name a metric is flagged.
That last rule is the Domino's moment, made routine: is this automation earning its place, or running because nobody turned it off?
The retainer also covers watching what is live: deliverability, failed actions and daily numbers. Anything wrong is raised the day it is seen.
The month-three test
Here is the test. At the end of month three, you should be able to answer yes to each of these.
- At least one automation built under the retainer is live and sending to real people.
- Each live automation has a goal set in Customer.io, added before it launched.
- You know your priority metric's baseline, the date it was taken, and where it stands now.
- The monthly report leads with that metric, not with opens and clicks.
- You can name what went live in each of the three months.
If you cannot, the retainer is buying meetings. That is not always the agency's fault: missing events or DNS records on the client's side can stall any agency. But the weekly update should say so plainly, with what is needed and from whom. Silence about a blocker is a delivery problem too.
The month-three test is a floor, not a ceiling. Proving an automation caused a lift takes more than a before and after. Customer.io's holdout tests send some people's messages to "an internal message trap", so you can compare conversions with and without the email (Customer.io docs). Our post on the Customer.io holdout checkbox shows how to set one up. A holdout needs volume, so use it on the busiest automations.
What a retainer costs
NerveCentral retainers run from £2,000/$2,900 to £12,000/$17,500 a month, depending on requirements (Customer.io agency). The difference is capacity: emails and automations a month, channels, and data work. Our Customer.io implementation cost guide breaks down what drives the number and how to compare quotes.
What a retainer excludes matters as much. Ours excludes three things:
- Work over the monthly capacity, unless agreed as extra.
- Your own development. We specify the events, attributes and DNS records; your team builds them.
- Customer.io fees and paid add-ons, which you pay Customer.io directly.
Ask any agency for this list in writing. Scope creep starts in the gap between what the client assumed and what the agency priced.
When a retainer is the wrong shape
A retainer suits a business that wants a live programme improved month after month. It is the wrong shape in four cases:
- You know exactly what you need built. A fixed set of automations with an end date is a build. We run those as 60-day or 90-day plans ending in a handover, with a recorded walkthrough of every live automation.
- Your data is not ready. If Customer.io is not receiving the events an automation would trigger on, start with the data plan, or an audit. Paying a monthly fee while you wait for engineering is the meetings retainer by another name.
- You have the skills in-house and lack only hours. A freelancer may suit you better. Our agency vs freelancer vs in-house guide sets out the trade-offs.
- You want one fix. One broken Liquid template is a small job, not a retainer.
Questions to ask before you sign
| Question | A good answer | A red flag |
|---|---|---|
| What will be live by the end of month one? | Named automations, with their trigger and goal | We start with discovery |
| Which one metric will you report first? | Your priority metric, with a baseline and a date | A dashboard of opens and clicks |
| When do you set the Customer.io goal? | Before the automation launches | We add tracking later |
| What is the monthly capacity? | A number of emails and automations, in writing | Whatever you need |
| What is not included? | A written list | A pause |
| What happens at month three if nothing has moved? | A review that stops, changes or replaces the work | These things take time |
A badge will not answer these questions for you. Our post on what the Customer.io certified partner badge proves explains why.
Want to put these questions to us? Our Customer.io agency and email marketing agency pages show how we work. Our case studies show the numbers. Then send us an enquiry or book a call with David.
Frequently asked questions
What does a Customer.io email marketing agency do each month?
A Customer.io email marketing agency on a retainer should put at least one named piece of work live each month: a new automation or a named change to a live one. Each is tied to one metric with a baseline. Calls, updates and reports support that work.
How much does a Customer.io agency retainer cost?
NerveCentral's Customer.io retainers run from £2,000/$2,900 to £12,000/$17,500 a month, depending on requirements (Customer.io agency). Clutch puts the average email marketing agency rate at $100 to $149 an hour (Clutch). Compare quotes on monthly capacity and what is excluded, not on the headline fee.
How long before a Customer.io retainer shows results?
A Customer.io retainer should have at least one automation live and measured against a baseline by the end of month three. Results take longer to prove where volume is low, but a live automation with a goal set before launch should exist by then.
What should a monthly email marketing report include?
A monthly email marketing report should lead with the business's priority metric against its baseline and target. Then it should show each automation with its numbers, what was done, what comes next month with the metric each piece moves, and what the agency needs from you.
How do I know if an email automation caused the result?
A holdout test is the most direct way to show an email automation caused a result. Customer.io's holdout test withholds a message from a share of people and compares their conversions with those who got it (Customer.io docs). It needs enough volume to reach a clear result, so use it on the busiest automations.
Should I choose a retainer or a fixed-term build?
Choose a fixed-term build when you know which automations you need and want an end date. Choose a retainer when you want a live programme reviewed and improved every month. A retainer can open with a 90-day plan and then move to monthly planning.
Do Customer.io partners offer retainers?
Yes, Customer.io's partners page says its partner team helps agencies "grow service revenue with implementation and retainer services" (Customer.io partners). Partner status does not tell you what a retainer will ship, so ask each partner what goes live in month one and how they will report it.
Sources
- Nation's Restaurant News, Domino's do-over, read 16 October 2026.
- Nation's Restaurant News, Domino's new pizza delivers sales surge, read 16 October 2026.
- Domino's Pizza, first quarter 2010 results release, 4 May 2010.
- Clutch, Email Marketing Pricing Guide 2026, read 16 October 2026.
- Customer.io, Partners, read 16 October 2026.
- MarketingProfs, Marketing Agency Pricing and Cash Flow Trends for 2025, read 16 October 2026.
- Ignition, 2025 Agency Pricing and Cash Flow Report, read 16 October 2026.
- Setup, The Truth About Agency Breakups, sixth Marketing Relationship Survey, released late 2024, read 16 October 2026.
- Customer.io docs, Automation goals and conversion criteria, read 16 October 2026.
- Customer.io docs, Holdout tests, read 16 October 2026.
- NerveCentral, Slimfy case study.
- NerveCentral, Customer.io agency, read 16 October 2026.
Want this working in your Customer.io workspace?
It's what we do all day for SMBs—strategy, automations, deliverability and hands-on execution.
See how we work as a Customer.io agency →
