You’ve got two versions of a landing page. Or two email subject lines. Or two ad creatives. You know you should test them against each other before picking a winner. But then you look into A/B testing tools and it’s a wall of enterprise pricing, JavaScript snippets, and statistical significance calculators.
What if I told you the only thing you actually need is two short links?
No platform. No SDK. No $200/month subscription. Just two tracked URLs, a way to split your audience, and five minutes.
The Simplest A/B Test You Can Run
Here’s the core idea: you create two versions of whatever you’re testing. You create a short link for each version. You send half your audience to Link A and half to Link B. Then you check which link got more clicks.
That’s it. That’s the whole framework.
It’s the same logic that Optimizely, VWO, and every other testing platform uses under the hood. They route traffic to different variants and measure which one performs better. The difference is they automate the splitting, add statistical models, and charge you for it.
For most decisions you’ll make — which headline to use, which CTA works, which email drives more visits — you don’t need automated splitting or p-values. You need directional data. And two tracked links give you exactly that.
Setting It Up: Five Minutes, Two Links
Let’s walk through a real example. Say you have two versions of a landing page and you want to know which one converts more visitors into signups.
Step 1: Create Your Two Variants
You need two distinct URLs. These could be:
- Two different landing pages (
/pricing-v1and/pricing-v2) - The same page with different query parameters (
?headline=aand?headline=b) - Two completely different domains if you’re testing brand positioning
The key is that each URL represents one version of whatever you’re testing.
Step 2: Create Two Short Links
Head to gofwd.to and create a short link for each variant.
On the free plan, you’ll get random short codes — something like gofwd.to/s/x7kQ2 and gofwd.to/s/m9pL4. These work perfectly fine.
On the Pro plan, you can use custom codes that make your life easier:
gofwd.to/s/promo-a→ your first landing pagegofwd.to/s/promo-b→ your second landing page
Custom codes aren’t just vanity — they make it dead obvious which link is which when you’re checking results later. When you’re staring at your analytics at 11pm trying to remember which random code maps to which variant, you’ll wish you had named them.
Step 3: Split Your Audience
This is the part where enterprise tools earn their money — they auto-split traffic for you. You’ll do it manually, and that’s fine.
A few ways to split:
- Email campaigns: Send Version A’s link to the first half of your list, Version B’s link to the second half. Most email tools let you segment by even/odd subscriber position or random split.
- Social media: Post with Link A on Monday, Link B on Wednesday. Or Link A on Twitter, Link B on LinkedIn (if you’re testing cross-platform).
- Ads: Run two identical ad sets with different destination links.
- Team distribution: Give Link A to your sales team and Link B to your marketing team to share.
The important thing: try to keep the audiences comparable. Sending Link A to your most engaged subscribers and Link B to dormant ones will poison your results.
Step 4: Let It Run
This is the hardest part — doing nothing.
Give your test at least 48 to 72 hours. Ideally a full week if you can stomach it. You want enough clicks on both links to see a pattern, not just noise from a handful of early visitors.
Don’t peek at results after two hours and declare a winner. That’s like calling an election with 3% of votes counted. Let the data accumulate.
Step 5: Check Your Analytics
Log into gofwd.to and look at the click data for both links. You’ll see:
- Total clicks — the headline number
- Clicks over time — did one variant start strong then drop off?
- Geographic data — are different regions responding differently?
- Device breakdown — mobile vs. desktop performance
- Referrer data — where clicks are actually coming from
Compare the two links side by side. The story is usually in the total clicks, but the breakdown data can reveal surprises. Maybe Variant A crushed it on mobile but Variant B won on desktop. That’s useful information that a simple “A vs. B” comparison would hide.
What You Can Actually Test
The two-link approach works for anything where you can create two versions and split an audience. Here are the most practical use cases:
Landing page headlines. Create two pages with different headlines, everything else identical. This is the classic A/B test and it works beautifully with short links because the tracking tells you exactly how many people arrived at each version.
Email subject lines. Split your list in half. Same email body, different subject lines, different short links in each. The link with more clicks had the better subject line. Simple.
Call-to-action variations. “Start Free Trial” vs. “See It In Action.” Different CTA, different link. Whichever gets more clicks is the one people actually want to click.
Ad creative. Running the same offer with two different images or copy angles? Give each one its own tracked link. Your ad platform’s analytics will show impressions and clicks too, but having independent tracking gives you a second source of truth.
Social media posting times. Same content, same link destination — but post with Link A at 9am and Link B at 2pm. Which time slot drives more clicks? Now you know when your audience is actually paying attention.
QR code placements. If you’re using QR codes for your restaurant menus or event materials, create two short links behind two QR codes and place them in different locations. The counter on the table vs. the one on the door. Now you have real data on which placement gets scanned more.
Reading the Results
You don’t need a statistics degree to interpret your results. Here’s the practical version:
55/45 split = noise. If Link A got 55 clicks and Link B got 45, you haven’t learned much. That’s close enough to be random variation, especially with small sample sizes. Don’t make decisions based on a 10% difference with under 200 total clicks.
70/30 split = signal. If Link A got 70 clicks and Link B got 30, now you’re seeing something real. That’s a clear preference, even with relatively small numbers.
What about the middle? A 60/40 split with 500 total clicks is probably meaningful. A 60/40 split with 50 total clicks is probably not. More data makes smaller differences more reliable.
Let’s make it concrete. You’re testing two email subject lines:
- Subject A (“Your free report is ready”) →
gofwd.to/s/report-a→ 312 clicks - Subject B (“I found something you’ll want to see”) →
gofwd.to/s/report-b→ 187 clicks
That’s a 62/38 split across 499 total clicks. Subject A clearly resonated more. Use it for the rest of your list, use it as a template for future emails, and move on.
The goal isn’t academic precision. It’s making a better decision than flipping a coin.
The Honest Caveat
I should be straight with you about what this is and what it isn’t.
This is not a statistically rigorous experiment. You’re not controlling for every variable. You’re not calculating confidence intervals. Your sample sizes are probably too small for a data scientist to take seriously. The way you split traffic introduces bias — people who open email on Monday morning are different from people who open it Wednesday afternoon.
And none of that matters for most decisions you’re making.
Here’s why: the alternative isn’t a perfectly controlled experiment. The alternative is guessing. Picking a headline because it “feels right.” Going with Version A because the boss likes it. Running with whatever you made first because testing feels like too much work.
Directional data beats no data every single time. If your short link data says 70% of people preferred Version A, you might not know the true preference with academic certainty — but you know a lot more than you did before. And it cost you zero dollars and five minutes.
The real risk isn’t that your test is imperfect. It’s that you never test at all.
When to Upgrade
There comes a point where two short links won’t cut it. If you find yourself in any of these situations, it’s time to look at dedicated A/B testing tools:
- You need automatic traffic splitting — sending all visitors to one URL and having the tool randomly show different variants
- You need statistical significance calculations — actual confidence intervals and sample size estimations
- You’re running multi-page funnel tests — testing entire user flows, not just individual pages
- You’re testing on-page elements like button colors, layout changes, or form fields without creating separate URLs
- Your decisions carry high financial stakes and a wrong call costs more than the testing tool
Tools like Optimizely, VWO, or even Google Optimize (while it lasted) exist for good reasons. But if you’re a solo founder, a small marketing team, or someone who just wants to make slightly better decisions slightly more often — you’re not there yet. Start with two links.
Start Your First Test
You now know everything you need to run an A/B test without paying for another SaaS subscription.
Pick something to test. Anything. Two headlines, two images, two CTAs — whatever’s been bugging you. Create your variants, grab two short links from gofwd.to , split your audience, and wait a few days.
Two links. One decision. Five minutes.
The best test is the one you actually run. Go run yours.