Results are called at 95% confidence and every losing test stays in the log beside the wins. Before anything starts, we check whether your traffic can support testing at all. The same people who write the hypothesis build the variant.
Head office in Mississauga, serving clients across Canada
How does CreateITivity handle CRO?
Win more customers from the visitors you already have, using research and testing at each step from landing page to checkout.
4 cro services to choose from
Monthly fees linked to targets you agree in writing
Fees: Custom quote one-time → Custom quote /month
Free 30-minute first call with the person who would lead your account
Our commitments
What you can expect from our CRO team
Anyone can change a page and show a graph that rose afterwards. Testing becomes a discipline when there is a firm rule for calling a result, and an honest record of the tests that did not win.
What counts as a result
A 95% significance level, measured at a sample size we agree on before launch. We never end a test early because it looks good, since that is how most inflated lifts get reported.
When a test loses
It is logged next to the winners, together with the idea that proved wrong. Losing tests are normal. A track record that only shows wins has been filtered.
Who approves the hypothesis
You sign off on every idea before we build it. Nothing launches on a hunch you have not seen, and the reasoning is written out for you to read.
When we will tell you not to test
When your volume is too low to reach an answer. The planner above runs the same maths we use. If it shows months, the audit will say so, and we will suggest other work instead of a testing retainer.
No lock-in
We use whatever testing tool you already pay for. Nothing proprietary disappears when the contract ends: the variant code, the results and the tracking all stay in your accounts. Who builds the variants
What it costs
Quoted in Canadian dollars once we know your traffic, because volume decides how many tests are worth running. How pricing works
Where to start
How conversion problems usually show up
People rarely arrive saying the conversion rate is the issue. They arrive with one of these three, and the cause sits underneath.
Does your page get enough traffic to test anything?
A split test needs volume to reach a verdict. Too few visitors and it simply runs out the clock with nothing to show. Enter your page's figures below and the calculator works out how long one honest test would take.
8weeks until the test can be called
012 weeks
Total sample, control plus variant
106,416
Tests you could finish per year
6
You can test, a few times a year. With so few slots, picking what goes first is the real decision. An audit up front helps you spend those slots on the changes most likely to matter. Book a CRO audit
How it is calculated: the standard two-proportion sample size, 95% significance (two-sided), 80% power, a single variant against the control. We never call a test in under two weeks, since weekends and weekdays behave differently. The result estimates the run time only. It cannot predict which version wins.
Your platform’s limits
The platform sets the limits on testing long before the budget does
Some test ideas die the moment you check what the platform allows. Here is what each stack we work on lets you change, and what it locks.
Hosted platforms
Quick to test on most pages, with a few areas you are not allowed to touch.
Product and collection templates are open to testing. On a standard plan the checkout is not, so the shipping-cost surprise that loses Canadian shoppers sits outside your reach.
Almost everything can be tested, and almost everything can break. Page builders and plugins rewrite markup on the fly, which makes client-side variants flicker or fail without warning.
Tidy markup means visual tests apply without fuss. With no server of your own, anything that must be decided before the page draws, such as showing French copy to Quebec visitors, is hard to do.
Flags live on the server, which lets a test change prices, the steps in a sign-up or where a visitor is sent next, well beyond colours and copy.
How it runs
A 90-day loop that restarts from a higher base
Tracking gets fixed before any test, since a bad measurement makes every result suspect. On busy pages the first proven winner usually arrives in 30–45 days, as the FAQ says. With less traffic, the planner above gives the realistic figure for your site.
Plenty of reported conversion wins are random variation with a convincing percentage attached. If a test has too few visitors to detect the change you hope for, stopping it at the first good-looking day is closer to tossing a coin. Before launch we work out the smallest effect worth detecting and the sample needed to see it. The test then runs until it gets there, and only a change the numbers support goes live on your site.
Sample size and statistical power are fixed before launch
No early peeking, since checking too often multiplies false winners
Traffic split checks flag broken randomization within days
Losing tests are shared in full because they shape the next idea
02
Test calendar
One change per test, decided at 95% confidence
Gains that stack
Small proven improvements add up to a much larger result
One modest improvement on a checkout page barely shows up in the monthly numbers. A year of modest improvements, each proven and each building on the same visitors, changes the business. The real value sits in the routine of measuring, diagnosing, testing, shipping and checking again. A Winnipeg retailer that keeps that rhythm for two or three quarters will usually see revenue per visit move further than a one-off redesign would manage.
Each win builds on traffic you already pay to attract
A regular testing rhythm outperforms the occasional full redesign
Each released change is measured again against what it predicted
What we learn goes back into the plan so ideas improve
Weigh it first
Do you have a conversion problem or a traffic problem?
Testing improves the share of visitors who act. With only a few hundred visits a month, even a big lift is a handful of leads. Settle where the money goes first.
Short answers on cost, timing and how we work with your team.
How is CRO priced?
Most clients start with a one-time CRO audit that produces a ranked list of conversion problems and a 90-day test plan. Ongoing work runs as a monthly retainer that covers test design, variant builds, analysis and weekly reporting. Both are priced to scope, and higher test volumes are available for larger sites.
How much traffic do I need for CRO testing to work?
A/B tests need a healthy number of conversions each month, often around a thousand across the funnel, to reach a reliable result in reasonable time. If you are below that, we rely on session recordings, heatmaps and customer interviews to make improvements without waiting for statistics. The audit tells you plainly which route suits your volume.
How long before I see a lift in conversions?
Pages with the most traffic reach a result first, so early wins tend to come from there. Broader gains across the whole funnel build over several months. Sites with less traffic take longer, because each test needs a big enough sample before we can call it.
What exactly is included in a CRO retainer each month?
An experienced conversion strategist, a ranked backlog of experiments, design and front-end build for each variant, statistical analysis, a written record of what we learned and a live dashboard. Nothing goes live until you approve the hypothesis behind it.
Do you run A/B tests only, or other types too?
We run A/B tests, multivariate tests and split-URL tests, plus server-side tests for personalization on larger programs. The data decides the method. Losing tests are written up as carefully as winners, so every result adds to what we know.
What tools do you use for testing and analytics?
We use whatever testing tool you already have, such as VWO, Optimizely or Convert, or suggest one that fits your stack and budget. Analytics runs on GA4 with server-side events, and behaviour data comes from Hotjar, Microsoft Clarity or FullStory. We avoid tools that would make you dependent on us.
How do you identify what to test first?
In the first week we set up GA4 events and session recordings and map each drop-off by the revenue it costs. Every problem becomes a hypothesis scored for potential, importance and ease. You sign off on the 90-day test plan before any variant is built, so we agree on priorities together.
Do you guarantee a conversion rate improvement?
Nobody can honestly promise a lift before seeing your data, since traffic quality, your offer and your pricing all play a part. What we commit to is the method: tests called at 95% confidence, honest reporting on wins and losses, and targets agreed in writing at the start.
What happens to the winning variants at the end of a test?
Winners go to your developers, or straight to production if we built them. Each one is logged with the evidence behind the decision. The next round of tests starts from the new baseline, so improvements stack.
Will CRO work conflict with our SEO or paid media programs?
When done well it helps both. A landing page that converts better lowers your cost per lead on ads. Better engagement on the page supports SEO too. We schedule tests around big traffic changes from SEO or ads so results are not skewed.
What do we own at the end of the engagement?
Everything we produced: variant code, test results, the learning log and the analytics setup. There is no proprietary testing platform to lose. Your data stays in your GA4 account and your codebase.
Show us where visitors drop off and we will suggest what to test
Point us at the journey that matters and give a rough monthly volume. We reply with the first few things we would examine, whether you have the traffic to test, and what to do if you do not.
A personal reply within one business day
A clear answer on whether your traffic supports testing
Enter your URL and we go through the journey ourselves. You get the drop-off points we can spot and a frank view on whether your traffic is enough to test at all.
