Empty. Every month it stays empty is a month of guessing which version was better. Go shop the stack. Nothing is sent until you book.
Most stores start here →◆ That's the whole pack ◆
All five, one team, one bill. Goldie approves.
Starting at — final scope quoted on a call
Nothing gets charged — the button books a call.
Stop arguing about which version is better. We test it on your real shoppers, ship the one that wins, and write down what happened either way.
One test a month, with enough traffic behind it to trust the answer · losses reported like wins · no guaranteed-lift theater
from $1,500/mo
— research, one test built and run to a decision, and the write-up
The first test is usually live inside three weeks and decided four to six weeks after that. Three-month minimum — not to lock you in, but because one month can't finish a test, and we'd rather say that than sell you one.
What's in the box
SERVICE FACTS
| First 30 days | A funnel read of where shoppers drop off, a ranked list of what's worth testing, and the first test live |
| How often we test | One test cycle a month, run to a decision — not four half-finished ones |
| You'll need | Around 250 orders and 8,000 sessions a month. Under that, testing can't reach an answer and we'll say so |
| Core tools | Shoplift on your own Shopify account ($99/mo and up, billed to you), plus your analytics — you keep all of it |
| Who you get | A conversion strategist who builds the variants, with design support — in your Slack |
| What happens each month | The hypothesis before we build, and the result after — including the tests that lost |
| Bigger stores | Past roughly 750 orders/mo we test twice as often; past 2,000 orders/mo we test checkout conversion and pricing directly |
| Your work | Winners stay in your theme, the test log is yours, and nothing is locked to us |
Week 0
Find where they give up
Walk the funnel step by step and find the drop-off that's costing you the most.
RESEARCH
Weeks 1–2
Write it down first
What we think will happen, why, which number decides it, and how much traffic it needs. Agreed before we build.
HYPOTHESIS
Weeks 2–3
Build the variant
Built against your real theme and checked on phones, desktop and Safari before a single shopper sees it.
BUILDING
Weeks 3–7
Run it to a decision
It runs until it hits the sample we agreed on — not until it looks good on a Tuesday.
RUNNING
Ongoing
Keep the winner, log the loss
Winners get built in properly. Losers get written up. Both point at what to test next month.
DECIDED
product-page add-to-cart lift from one test — Modified Tot, Q1 2026
A single test, with the traffic to answer, run to a decision. The arithmetic below explains why it isn't four.
Every result written up the same way. Most tests don't win — you'll see those too.
Whether this is worth buying comes down to how big an improvement you're trying to spot. A test needs a certain number of orders in each version before the result means anything, and halving the improvement you want to detect roughly quadruples the traffic you need. Here it is for a store doing about 300 orders a month:
| To spot an improvement of | Orders needed in each version | How long that takes |
|---|---|---|
| 30% | ~175 | About 5 weeks |
| 20% | ~390 | About 2.5 months |
| 10% | ~1,570 | About 10 months |
+8%
more shoppers adding to cart from the product page, from a single test
Modified Tot · A/B test · Q1 2026
One change, on the page where the decision actually gets made. Add-to-cart was the decision metric precisely because it happens often enough to answer inside a month.
All the case studies →Big swings before button colors
At your traffic, only a large difference is detectable. So we test real changes — a different offer, a rebuilt product page, a different first screen — not tweaks that would need a year to prove.
The page where the money is decided
Product pages and cart before the homepage. Most stores get their homepage tested to death while the page that actually sells goes untouched.
Obvious breakage before clever ideas
If something is plainly broken on mobile, we fix it and move on. You don't need a four-week experiment to prove a broken form is bad.
★★★★★
After the rebuild and the testing program that followed it, Dura-Cleanse says their conversion rate roughly doubled — their words, and their number.
Dura-Cleanse Client-reported · rebuild and ongoing testing together, not testing alone
You're already paying for the visit. Testing decides how many of those visits turn into orders — the same traffic, worth more.
Testing finds the better version; Growth builds it in properly and keeps the store fast enough that the win survives.
They do, and the arithmetic doesn't support it at your size — the table above shows the work. Four tests a month at a few hundred orders means four coin flips reported as findings. We run one test properly, on a metric that has the traffic to answer inside a month.
Around 250 orders and 8,000 sessions a month, sustained. Below that a test cannot reach a decision in a sellable timeframe, and we will tell you so instead of selling you a retainer that can't finish. The honest option at that size is our Store Teardown — a one-time deep-dive CRO review with designed mockups — and implementation work. Our free conversion & AOV calculator shows what a lift would be worth at your volume.
Then you learned the change was wrong before you rebuilt your whole store around it, and you get the write-up anyway. Most well-designed tests do not win. An agency that only ever reports winners is not testing — it's shipping opinions and taking credit for whatever the season did.
No, and be careful with anyone who does. At a few hundred orders a month, a real 15% improvement and a lucky quarter look nearly identical from the outside — so a guaranteed lift is a promise about randomness. What we do guarantee is the process: a test built, run to its pre-registered sample target, and documented either way.
Shoplift, installed on your own Shopify account and billed to you directly — it starts at $99/mo and scales with your traffic. We don't mark it up or bury it in the retainer, and because it's yours, the account, the history and the test data stay with you.
No. The old generation of testing tools swapped content in the browser after the page loaded, which is what caused the flicker everyone remembers. Shoplift renders the variant against your theme before the page reaches the shopper, so there's nothing to flash.
Shopify Growth ships changes and keeps the store healthy. This decides which changes were actually better. They work well together and neither replaces the other — Growth builds the winner properly once a test has found it.
Every winner we shipped stays shipped — it's in your theme, not in a tool you're renting from us. The test log is yours, so the next person doesn't re-run the tests you already paid for. Cancel any time after the first three months.
GOLDEN DIGITAL
est. 2022 · Ecommerce growth, end to end.
SHOPIFY GROWTH · QTY 1from $2,250/MO
PAID ADS (META + GOOGLE) · QTY 1from $1,500/MO
KLAVIYO EMAIL & SMS · QTY 1from $2,000/MO
SEO & AI SEARCH · QTY 1from $2,000/MO
A/B TESTING & CRO · QTY 1from $1,500/MO
STORE BUILD (ONE-TIME) · QTY 1
TOTAL / MO
★ No fit? We'll say so on the call and send the notes anyway ★
goldendigital.co
Planning estimate only — nobody has charged anybody. Bring this to the call and we'll price the real thing.
Book the call → 30 min, free
What's in the box