The 3-Question Test: Is It Your Traffic, Your Page, or Your Offer?

Before you change a single thing on your store, answer three questions. They take about ten minutes, they use numbers you already have, and they tell you which layer of your funnel is actually broken, so you stop pouring effort into fixing the wrong one.

This is the direct antidote to the "we've tried everything" trap (Article 6). Tactics fail when they're aimed at the wrong layer. Diagnosis is just aiming first. By the end of this article you'll be able to look at three numbers and say, with real confidence, whether your problem is traffic, page, or offer, and exactly which article to read next to fix it.

The principle: diagnose the layer before you touch it

Almost every conversion problem lives in one of three layers:

• Traffic, are the right people even arriving?
• Page, once they arrive, are they convinced enough to add to cart?
• Offer, once they want it, are they persuaded to actually pay?

Each layer has a signal you can read today. And here's the rule that makes this work: walk them in order. A problem in an early layer makes the later ones impossible to judge, if the wrong people are arriving, it's meaningless to ask whether your page is "convincing" them. Diagnose top-down.

Q1, Are these the right buyers? (the traffic layer)

The signal: bounce rate, time on the product page, and how returning visitors behave.

If qualified, interested people are arriving, they engage, they scroll, they read, some come back. Healthy ecommerce bounce sits around 38–47%, and the best stores run 20–45% depending on category (Digital Web Solutions). If your bounce is far above your category norm and time-on-page is near zero, you likely have a traffic-quality problem, your ads are pulling in people who were never the right fit.

This is the layer everyone skips, and skipping it is expensive: no page or offer fix will rescue mismatched traffic. You can have the best product page on the internet and still "convert" terribly if the people landing on it were never going to buy. So rule traffic in or out first. If traffic looks healthy but sales still lag, good. That's your signal to move down a layer.

Q2, Are they convinced? (the page layer)

The signal: add-to-cart rate.

This is the cleanest single read on whether your page is doing its job. Right-fit visitors who understand and want the product add it to the cart. Shopify stores average an ATC rate around 4–4.6%; the top 10% clear 9.6%; and below roughly 5% is a recognized signal of a product-page problem, unclear value, weak proof, confusing options (UpCounting).

So the logic is simple: qualified traffic arriving (Q1 healthy) + sub-benchmark add-to-cart = the page isn't selling. That's where clarity, trust sequence, and decision flow come in (Articles 11, 15, 16). If you want to go deeper on why an ATC rate is weak, Article 10 breaks it into four ranked causes.

Q3, Are they persuaded to pay? (the offer layer)

The signal: add-to-cart → checkout completion, i.e. your cart-abandonment pattern.

If people add to cart but vanish before paying, the hesitation is about the deal, not the product. Cart abandonment averages ~70%, and the single biggest reason, 48% of abandoners, is unexpected costs at checkout (Triple Whale). Abandonment well above your category norm, especially cost-driven, points to the offer layer: price-to-value, shipping, guarantees, trust at the moment of payment.

One important nuance: most people start here, assuming abandonment is a checkout problem. It usually isn't, which is exactly why Article 9 argues cart abandonment is rarely a checkout problem at all. Q3 tells you whether the offer is the issue; it doesn't tell you to go bolt badges onto your checkout.

Reading them together

The power isn't in any one number, it's in the pattern. Walk the funnel top to bottom and find the first place the drop is bigger than your category norm:

• Bounce high, everything downstream unreadable → fix traffic first.
• Traffic healthy, ATC low → it's the page.
• Traffic and ATC healthy, checkout completion low → it's the offer.

You're looking for the earliest broken layer, because fixing it often clears the symptoms downstream too.

A worked example

Make it concrete. Say a store has a 41% bounce (healthy for its category), a 2.8% add-to-cart rate (well below the ~5% line), and a 68% cart-abandonment rate (right around normal). Walk it: traffic is fine, abandonment is fine, but add-to-cart is sub-benchmark. Diagnosis: it's the page. The right people are arriving and most who add to cart finish, but far too few are getting convinced enough to add at all. This store should be working on clarity and trust on the product page, not running checkout-friction fixes or buying more traffic. Three numbers, ten minutes, one clear direction.

"But what if it's all three?"

Sometimes it genuinely is, and I won't pretend the framework dissolves that. When multiple layers are weak, the three-question test doesn't tell you to fix everything at once; it tells you which to fix first, which is its own discipline (that's Article 18, prioritizing by revenue impact). The framework guides; it doesn't dictate.

One honest caveat: at lower traffic, these metrics get noisy, the same low-volume problem from Article 6. Read them over a meaningful window, not a noisy week, and weight the layer where the gap to benchmark is largest and most consistent.

How to pull the three numbers in ten minutes

So this isn't abstract, here's where each lives:

• Bounce / engagement, GA4 (Engagement → Pages and screens), or your analytics app's landing-page report.
• Add-to-cart rate, Shopify Analytics (Online store conversion over time) shows "added to cart"; divide by sessions.
• Checkout completion / abandonment, Shopify's conversion funnel (reached checkout → completed), or your checkout/abandonment report.

Write the three numbers down next to your category benchmarks. The biggest gap is your diagnosis.

A second worked example (the offer layer)

Here's a different pattern, so you can see the framework flex. A store has a 39% bounce (healthy), a 6.1% add-to-cart rate (above benchmark, the page is clearly selling), but a 79% cart-abandonment rate, and the abandonment data skews heavily toward "shipping too expensive." Walk it: traffic is fine, the page is great at convincing people to want the product, but they're bailing at the deal. Diagnosis: it's the offer, specifically cost/shipping. This store should not touch its product page; it should fix shipping pricing, surface costs earlier, or rework the offer math (Article 8). Same three numbers, completely different prescription from the first example.

Why the order matters more than any single metric

It's tempting to jump to whichever number looks worst. Resist that. A scary-looking abandonment rate is meaningless if your add-to-cart rate is already tiny, because the few people reaching checkout aren't a representative sample, and "fixing" checkout would polish a step almost nobody reaches. The top-down order exists precisely to stop you from optimizing a downstream layer while an upstream one is quietly poisoning every number beneath it.

Common mistakes when running this test

• Reading a noisy week. Use a 30-day window, not 7 days, especially at lower traffic.
• Ignoring device and source splits. A "fine" blended number can hide a broken mobile or cold-traffic segment (Article 13, Article 1).
• Comparing to the global average instead of your category. Food & beverage and luxury live in different universes, benchmark against your own.
• Stopping at one layer. Sometimes the earliest broken layer isn't the only one. Fix it first, then re-run the test.

When three questions aren't enough

The 3-question test is a triage tool, it tells you which layer to work on, fast, with numbers you already have. What it doesn't do is tell you the specific fix within that layer. If it points at the page, you still need to know whether the issue is clarity, trust, or decision flow. That's the next level of diagnosis, and it's exactly what the depth articles (and the Scorecard) are for. So think of this test as the fork in the road, not the whole map: it gets you pointed in the right direction in ten minutes, which is most of the battle, because the most expensive mistake in CRO is confidently fixing the wrong layer.

Key takeaways

• Every conversion problem lives in one of three layers: traffic, page, or offer.
• Diagnose top-down, an early-layer problem makes later layers unreadable.
• Signals: bounce (traffic, ~38–47% healthy), add-to-cart rate (page, sub-5% = problem), checkout completion / abandonment (offer, ~70% avg).
• Read the pattern, not one number, fix the earliest layer below benchmark.
• If it's all three, that's a prioritization question (Article 18), not a reason to fix everything at once.

Where to go next, based on your answer

• Traffic problem? Start with who you're targeting and why "good traffic" stops converting (Article 1).
• Page problem? Weak add-to-cart (Article 10), the silent questions buyers ask (Article 11), the 5-second test (Article 15), trust as a sequence (Article 16).
• Offer problem? When the offer is the real issue (Article 8).
• Cart abandonment specifically? Article 9.