Testing and Optimization: Real Results
Small changes create big wins. See how A/B testing works in practice, plus the metrics you should actually track instead of vanity numbers.
Why Most Companies Test Wrong
You’re probably tracking the wrong metrics. Bounce rate. Page views. Time on site. These numbers feel important, but they don’t tell you if anyone’s actually converting. We’ve watched dozens of Canadian businesses pour effort into optimization and miss the real opportunities because they’re measuring noise instead of signal.
Testing isn’t complicated. It’s actually pretty straightforward — you change one thing, measure what matters, and decide if it worked. But here’s what we’ve learned: most teams either don’t test at all, or they test everything at once and can’t figure out what actually made the difference.
The Testing Framework That Actually Works
Testing isn’t about luck. There’s a process. Follow it, and you’ll see results. Ignore it, and you’re just guessing.
Identify Your Bottleneck
Where’re people dropping off? Not your hero section if your conversion rate is 0.8%. It’s probably later in the funnel. Use your analytics to find the step where the most people bail out.
Form a Hypothesis
Don’t just guess. “I think the CTA button should be bigger” isn’t enough. It’s “I think people aren’t clicking the form button because it blends into the background. If we make it stand out with a contrasting color, we’ll see 15-20% more clicks.”
Run the Test (Properly)
50% of traffic sees the original, 50% sees the variation. Run it for at least 2 weeks or until you’ve got 200+ conversions in each version. Statistical significance matters. A 5% improvement on 20 conversions? That’s noise.
Learn and Repeat
If it won, implement it and test something else. If it lost, document why and move on. You’re not trying to find one magic change—you’re building a process where each iteration gets slightly better.
Metrics That Matter (And Ones to Ignore)
Not all numbers are created equal. Some actually tell you if your page is working. Others just look impressive in a report.
Conversion Rate
Visitors who complete your goal (sign up, download, purchase) divided by total visitors. This is your north star. Everything else is supporting evidence.
Cost Per Conversion
If you’re paying for traffic, this tells you if your page is profitable. A $5 conversion cost on a $50 product is terrible. On a $200 service? That’s solid.
Checkout Abandonment Rate
For e-commerce, this is where you’ll find your biggest quick wins. If 70% of people start checkout but don’t finish, that’s your priority. Not traffic volume.
Bounce Rate
A 40% bounce rate could mean you’re attracting exactly the right people who find your answer immediately and leave. Or completely the wrong people. You can’t tell from this number alone.
Time on Site
Someone spending 10 minutes reading might be thoroughly confused. Someone spending 2 minutes might have converted and bounced happily. This metric doesn’t measure anything useful for conversion.
Pageviews
You could have a million pageviews and zero revenue. It’s the definition of a vanity metric. Focus on what people do, not how many pages they see.
A Real Example: From 1.2% to 2.8% in 8 Weeks
We worked with a Toronto-based SaaS company that was stuck. Their landing page had been the same for almost a year. Conversion rate? 1.2%. That’s not terrible, but it’s not great. They weren’t sure where to start.
Week one, we just watched. Heatmaps showed people scrolling past the main value prop and spending time reading the feature list. The form had 5 fields and people were abandoning it about 60% of the time. The hypothesis was simple: fewer fields, more conversions.
We cut the form down to 3 required fields (name, email, company size) and moved the rest to a follow-up. Conversion rate jumped to 1.8% immediately. Not huge, but it validated the direction.
Next, we tested the hero headline. Original was generic: “Enterprise SaaS for Growing Teams.” We changed it to speak to their actual pain point based on customer interviews: “Stop Losing Track of Who Said What.” Conversion hit 2.1%.
Third test was the CTA button. They’d been using gray. We tested a contrasting teal button with “Get Started Free” instead of “Sign Up.” That got them to 2.5%.
The final test that moved the needle was removing the pricing table from the page and replacing it with “Let’s Talk Pricing” — a button that opened a calculator. People who saw actual numbers were overthinking. This change? 2.8% conversion.
“We didn’t need a redesign. We needed to stop guessing and actually listen to what our visitors were telling us through their behavior.”
— Marketing Director, SaaS client
That’s 133% improvement in 8 weeks. No new traffic source. No major design overhaul. Just one small test at a time, based on data, not opinions.
Common Testing Mistakes to Avoid
You can do everything right except one thing and still waste weeks on a failed test.
Testing Too Many Things at Once
Change the headline, button, and form all at the same time? Now you don’t know which one mattered. Test one variable. That’s it.
Not Running Long Enough
Monday you’re at 3%, Wednesday you’re at 2%, and you kill the test. But that’s noise. You need 2+ weeks minimum, or 200+ conversions per variation — whichever comes first.
Optimizing for the Wrong Metric
You increase clickthrough rate but conversion rate stays flat? You’re attracting the wrong people, not the right ones. Always test against your actual goal.
Ignoring Seasonality
Your page converts great in December but tanks in July? That’s normal for most industries. Don’t compare results across seasons. Test during similar periods.
Tools You Actually Need
You don’t need fancy software to run tests. But you do need something that gives you statistical confidence.
Google Optimize (Free)
If you’re already in Google Analytics, this integrates directly. Simple to set up. Works fine for straightforward tests on landing pages.
Unbounce (Paid)
Built specifically for landing page testing. No coding required. You build the variation in their editor and it automatically splits traffic. Good for non-technical teams.
VWO (Paid)
Visual Website Optimizer. Heatmaps, session recordings, and A/B testing all in one. More expensive but gives you the context you need to understand why changes work.
Hotjar (Paid)
The heatmaps and recordings here are exceptional. You’ll see exactly where people click, scroll, and abandon. Use this to form hypotheses before you test.
Crazy Egg (Paid)
Another solid heatmap tool. Similar to Hotjar. The real value is the scroll maps that show where people actually look on your page.
Your CRM (Essential)
Whatever system you use to track customers—HubSpot, Salesforce, Pipedrive—integrate it with your testing tool. You need to know which variation produces better-quality leads, not just more leads.
Start Testing This Week
You don’t need to wait for perfect data or a complete redesign. Pick one page that’s underperforming. Look at your analytics and identify where people are dropping off. Form one hypothesis. Run one test. See what happens.
Small changes compound. A 1% improvement this month, 1.5% next month, 2% the month after. In six months you’re 20% better. That’s the power of testing.
Ready to Optimize Your Pages?
We work with Canadian companies to run landing page tests that actually move the needle. If you’re not sure where to start, let’s talk about what your data is telling you.
Get in TouchImportant Note
This article is informational and educational in nature. The examples and case studies mentioned are based on industry practices and real-world scenarios. Your results will vary based on your specific audience, industry, and current conversion funnel. We recommend consulting with conversion optimization professionals before implementing significant changes to your landing pages. All metrics and statistics discussed represent typical scenarios and shouldn’t be interpreted as guarantees for your business.
Related Articles
