Hypothesis
If search clicks go to a dedicated landing page that matches the query instead of an existing site page, more clicks will convert, because the page answers the searched intent without asking the visitor to navigate. The decision behind it is concrete: send search ad clicks to a dedicated landing page or to an existing product, category or home page.
The hypothesis is not as safe as it sounds. A dedicated Landing PageThe page a visitor arrives on from a campaign, built to turn that visit into one action. can match the ad's promise exactly and remove distractions, but an existing product or category page can offer choice, reviews, navigation and trust signals that a stripped-down page lacks, and some searchers want to browse. The landing page versus homepage guide sets out when each tends to suit; this test settles it for one campaign with real traffic. A dedicated page also has a running cost: it has to be built, kept in line with prices and stock, and tracked, so it must earn more than the existing page to be worth keeping.
Metrics and guardrails
The primary metric is conversions per click, the paid-traffic conversion rateShare of visits or clicks that convert. of each arm. Read it from the campaign experiment's report, or from analytics filtered to the experiment's traffic, so that both destinations are judged on the same visitors: ad clicks only, not organic or direct arrivals who happen to land on the existing page.
Two guardrails keep the win honest. The first is average order value for a shop, or lead qualification rate for a lead business: a focused page can raise conversions per click by narrowing choice, and if the orders are smaller or the leads weaker, the gain is less than it looks. The second is the landing page experience component of Quality ScoreGoogle's 1–10 keyword diagnostic.. Google's Quality Score documentation lists landing page experience as one of its components, and the Quality Score guide explains how it affects what each click costs. A destination that converts better but rates worse may lose some of its gain through higher costs, so read both.
Design
Use the platform's campaign experiment with the final URL as the only difference. Duplicate the campaign or ad group as a trial arm, change the final URL of every ad to the dedicated page, and leave keywords, ads, bids, budgets and schedules identical. Users are split between the arms by the experiment's traffic split. Before launch, click through both arms, submit a test conversion on each destination and confirm that both record conversions the same way.
| Design element | Choice | Reason |
|---|---|---|
| Unit of randomisation | User, through the campaign experiment's traffic split | The ad and auction stay identical; only the destination differs |
| Arms | Existing site page versus a dedicated landing page | One change: the final URL |
| Ads and keywords | Identical in both arms | Isolates the destination from the message |
| Tracking | Same conversion actions, tested on both destinations before launch | Unequal tracking produces a false winner |
| Primary metric | Conversions per paid click | Judges both pages on the same visitors |
| Promotions | None that touch only one destination | A sale on one page decides the test by itself |
Write the dedicated page to match the queries the campaign serves, the principle the message match guide explains; Message MatchHow closely a landing page repeats the promise and wording of the ad or link that sent the visitor. is the reason the hypothesis expects a gain, so a dedicated page that does not match the ads tests nothing useful. Keep the offer, price and terms identical on both destinations.
Sample size reasoning
Clicks are plentiful but conversions are not, so size the test on the conversion rate of the current destination and the smallest relative gain worth building and maintaining a separate page for. A small gain may be real and still not pay for the page, so decide that threshold first. The sample size and significance guide explains how the baseline rate, the minimum gain and the required confidence combine into a sample per arm, and the conversion rate calculator set up for landing page tests turns that gain into enquiries or orders so the threshold can be set in business terms.
If the campaign cannot reach that sample in a tolerable period, test on a larger campaign whose answer would guide the smaller ones, or test a bigger difference between destinations whose effect would be larger.
Duration
Run for at least two full weekly cycles, longer where the conversion window is long, and fix the end date before launch. Never run across a promotion, a price change or a stock change that touches only one destination. If the existing page is edited by another team during the test, record the change and, where it is large, restart the clock.
Reading the results and the decision rule
Read the test once, at the planned end, after the conversion window has passed. Compare conversions per click between arms with the confidence interval, then check order value or lead quality and the landing page experience rating. Write the decision rule into the plan before launch:
| Result | Decision |
|---|---|
| Dedicated page wins on conversions per click, order value or lead quality holds | Adopt the dedicated page for this campaign |
| Dedicated page wins on conversions but loses on order value or lead quality | Keep the existing page, or revise the dedicated page and retest |
| Existing page wins | Keep it and improve it; the site's navigation and trust signals are doing work |
| Difference inside the noise | Keep the page that costs less to maintain, usually the existing one |
Once the test has finished and both arms' visitors and conversions are known, the significance reader below shows how confident you can be that the difference between the two conversion rates is real. It reads a finished two-arm test; it does not tell you how many visitors to plan for.
Significance reader for a finished destination test
A kitchen fitter tests a dedicated page for its design consultation ads
Hollowmead Kitchens, an invented kitchen design and fitting company, sends its consultation ads to its home page. It builds a dedicated page that repeats the ad's offer of a free in-home design visit and runs a campaign experiment with an even split for five weeks. The home page arm records 3,900 clicks and 117 consultation requests; the dedicated page arm records 3,850 clicks and 158 requests, conversion rates of 3.0% and 4.1%. The sales team's qualification rate is similar in both arms, and landing page experience is rated the same. Hollowmead adopts the dedicated page for consultation campaigns.
Pitfalls
Different tracking on the two destinations is the most damaging pitfall, because it produces a confident and false winner. A dedicated page built on a separate tool may fire the conversion tag twice, or not at all on some browsers; test both before launch and compare conversion counts against the business's own records after. Site navigation leaking traffic is the second: visitors on the existing page may wander to other pages and convert there, which counts for that arm only if the conversion is tracked site-wide. Make sure both arms count conversions wherever on the site they happen.
Organic visitors on the shared page are the third. The existing page's overall figures include people who arrived from search results, email or bookmarks, so never compare its general conversion rate with the dedicated page's paid-only rate. If clicks convert but the account still shows few conversions, the clicks but no conversions troubleshooting guide helps separate a page problem from a tracking or traffic problem.
Common mistakes
Destination tests fail on their setup more often than their statistics. The landing page headline test changes one element on one page and is the natural follow-up once the destination is settled, and the experiment library lists every design.
- Changing the ads with the destination. The message becomes a second variable.
- Tracking the two pages differently. A tag that double-counts decides the test.
- Comparing against the page's site-wide rate. Organic visitors are not paid clicks.
- Ignoring order value or lead quality. A focused page can win on count and lose on value.
- Building the dedicated page without message match. It then tests design, not relevance.
- Using the significance reader to plan the test. It reads a finished test; plan with the sample size guide.
