What CRO costs, and what it returns
For a store doing $1–10M a year, budget $60,000–$180,000 all-in for a year of serious CRO: the retainer, the testing tool, and enough developer time to ship what wins. A well-run first year returns roughly 10–15% more revenue, not a doubling.
Almost every page on the internet titled “how much does CRO cost” refuses to give a number. This one will.
Two reasons the honest version is worth writing. First, most of the buyers who need this are $1–10M ecommerce founders who do not have the time to sit through five sales calls to build a range from silence. Second, we are the agency writing it, so we would rather qualify people at the top of the funnel than book calls we cannot help with.
The short version
For a store doing $1–10M a year, expect $60,000–$180,000 all-in for a year of serious CRO work. That range covers the agency retainer, the testing tool, and enough developer time to actually ship what wins.
The lower end assumes an audit plus a lightweight retainer. The upper end is a full testing programme with two to three tests running at any time, plus meaningful implementation work.
Below that spend, you are either buying a template or paying an agency that is under-servicing you. Above it, you are usually at a scale where an in-house team starts to compete.
The line items
Agency retainer: $4,000–$20,000 a month. This is the biggest number, and it is the one with the widest range in the market. Retainers under $3,500/mo are almost always either offshore, template-driven, or someone learning on your account. (Ours sits at $5,000 for the full plan, and month one is roughly break-even for us; the model only works because clients stay.) Retainers above $20,000/mo usually reflect either enterprise complexity (multiple stores, multiple markets, checkout extensibility work) or reputation pricing.
One-off audit: $5,000–$12,000. A structured research pass that produces a ranked problem list. Most agencies credit this against the first month of a retainer if you continue.
Testing tool: $0–$500 a month. VWO, Convert, Optimizely, and others. Free tiers exist and are enough for smaller stores. Shopify Plus’s native testing is included with Plus.
Analytics: usually free. GA4 is free. Shopify’s own reporting is free. Warehouse-scale setups (BigQuery, Fivetran) can add real cost but are not required to run a good programme.
Session recording tools: $0–$300/mo. Hotjar, Microsoft Clarity (free), FullStory. Do not need more than one.
Developer time: highly variable. This is the part that most cost quotes leave out, and it is often 30–40% of the total. Building variants, shipping winners properly, and cleaning up dead tests all cost developer hours. Budget $30,000–$60,000/yr for a moderately active programme, either on retainer or on your own team’s time.
Total for the year: roughly $60,000 at the low end, $180,000 at the top of a healthy $1–10M engagement. Enterprise programmes on $50M+ stores can run several times that.
What CRO actually returns
Here is the part that gets oversold most.
Realistic returns for a well-run first year on a $4M store: 10–15% total revenue lift. That is about $400,000–$600,000 in additional revenue. The math is in Chapter 1 of our course, and the shape is roughly:
- Roughly 1 in 5 tests wins, on a typical programme
- Winners average a 3–8% lift on the metric they targeted
- Compounding across four or five wins in a year, with normal interaction effects, lands most stores at 10–15% net
That is a very good year. It is not a doubling, and anyone promising a doubling is selling something else, usually a discount programme wearing a CRO label.
Return on the ROI calculator, applied honestly: a $4M store at 2.1% conversion running a proper programme should expect $400,000–$600,000 in year-one revenue from a $60,000–$120,000 investment. Positive ROI in nearly every case, but the returns come from research and shipping the obvious findings, not from clever tests.
Our ROI calculator will run these numbers on your actual traffic and AOV, with an honesty check on whether your traffic can even detect the lifts you are modelling.
Agency, in-house, or hybrid
Agency wins when: you do not have the internal expertise, you want breadth of experience across accounts, you value process that has been built over hundreds of tests, or the volume of work does not justify two full-time hires yet. Which is most $1–10M stores.
In-house wins when: you have the volume to keep two people busy year-round, your product needs deep context to test well, or you have found that agencies keep needing to be taught your business.
Hybrid works well at $10M+, where an in-house CRO lead runs strategy and reporting and an agency runs the testing programme. The lead keeps context; the agency brings breadth.
For a fully-loaded in-house senior CRO manager, budget $120,000–$180,000/yr plus tooling. A full team (manager, designer, developer) is $350,000+ before overhead. Which is why agencies exist in the first place.
The traps
“Guaranteed lift” agencies. Nobody can guarantee a lift on your specific store. Anyone who does is either lying or running a discount programme. Read Chapter 4 on how tests get accidentally rigged to always look like wins.
Testing-only agencies with no research phase. Testing without research means testing guesses. Roughly 1 in 5 guesses wins. That is a poor use of your traffic.
Redesigns dressed up as CRO. A redesign changes a hundred things at once. When the number moves, nobody can attribute it. When it drops, nothing can be rolled back. This is not testing.
Anyone charging less than $3,000/mo. For real work, that number does not cover the strategist’s time, let alone the tool, the developer, and the analysis. What you are getting is either a template or overseas work being resold.
When CRO is not worth the spend
Under about $1M in revenue, or under about 40,000 monthly sessions, a testing programme cannot resolve most changes fast enough to justify the retainer. What you should do instead:
- Run the free research method yourself, once, properly
- Fix the obvious findings and ship them without a testing programme
- Revisit CRO once you cross the traffic threshold
We will say this on the first call rather than sell you the retainer.
The honest counter: if your issue is a specific fixable thing (a broken checkout, a product page that fails on mobile, a landing page that does not match the ad), a one-off audit or engagement can absolutely pay back below $1M. It just should not be sold as an ongoing programme.
The commercial bit
We run CRO retainers and audits for eight to ten DTC brands at a time. Retainers land in the ranges above; the audit is fixed-fee. If you would like the honest read on which is right for your store, book a call or grab a free teardown first. Twenty minutes, no deck.
And if you are below the traffic threshold and we would waste each other’s time, we will say so. That is the entire arrangement.
Data-backed experiments, every Tuesday
3 short examples, 2 CRO tips, 1 teardown. Every Tuesday.