Promotion tool

Unbounce landing pages

Unbounce is a landing page builder. You design a page in the browser, publish it on a URL, and you can run an A/B test between variants of that page. It is a separate product from a separate company. Maxlead does not own Unbounce. Maxlead does not host your Unbounce variants, and Unbounce does not approve your CPA conversions. The page can be as carefully designed as you like. The call to action still has to be the Maxlead tracking link. It is never the advertiser preview URL.

That rule is the whole integration. Unbounce decides what the visitor sees. Maxlead decides whether the click was yours, whether the action was approved, and what the weekly payout includes. If the button points at the preview, Unbounce can report a healthy page and the network can report nothing. This guide explains who the builder is for, how to attach a CPA offer without breaking the link, a worked test, the mistakes that waste a variant, and how to read EPC and CVR on Maxlead before you send the winner more traffic.

What Unbounce is

Unbounce is a place to assemble a page without writing the layout by hand. You pick a structure, you write a headline, you place a button, and you publish. For a paid campaign, that page is the thing the ad promises. The visitor should land on a message that matches the ad and that matches the action the offer will ask them to complete. Unbounce is good at that presentation layer. It is not an offer catalog and it is not a payout system.

The A/B test is a split of your page, not a split of the advertiser. Variant A and variant B are both Unbounce pages you control. Visitors are divided between them. Unbounce can tell you which variant collected more button clicks, or more of whatever goal you configured inside the builder. A button click is not an approved CPA conversion. The goal inside Unbounce ends when the visitor leaves. The result that matters for earnings is recorded later, on Maxlead, after the tracking link has created a click and the advertiser’s action has been approved or rejected.

Maxlead is a CPA performance network and has operated since 2015. You get a publisher account, you are approved, and you run offers from that account. Public offers can be used once the account itself is approved. Other offers need a request and an admin decision. The tracking link is created for the offer and already includes your publisher id. You pass s1 through s8 for the source, the variant, the creative, and anything else you want back on the report. A global postback covers the account. An offer postback covers one offer and replaces the global URL there. Postbacks go out when a conversion is approved or rejected. Pending conversions do not send one. Approved earnings are paid every week. Pending and rejected rows are not paid as though they were approved.

The dashboard is where you read clicks, conversions, earnings, EPC, and CVR. Earnings means approved payout. CVR means approved conversions divided by clicks. EPC means earnings per click. Unbounce does not calculate those figures for you, because Unbounce does not know which of your button clicks became approved actions. If you scale the variant Unbounce liked, and you never look at EPC on Maxlead, you can scale a page that produces clicks and rejections.

Who it is for

Unbounce fits a publisher who wants a dedicated page for paid traffic and does not want to maintain a template in code. Native buyers, search buyers, and social buyers use it when the ad needs a specific URL that is not the advertiser’s domain. You can change a headline, republish, and send the next click to the new copy without asking a developer to deploy.

It is the wrong tool if the offer forbids a landing page and requires the ad to hit the tracking link directly. Building a beautiful Unbounce page against that rule does not make the traffic allowed. It is also the wrong tool if you only needed a banner. A page builder does not draw the image the native network serves. And it is the wrong place to look for offers. The offer lives on Maxlead. Unbounce only holds the page that points at the link.

Publishers who are new to the account should get one offer approved before they design three variants. The country list and the allowed sources decide whether the Unbounce URL is even usable. An account manager on Maxlead can confirm that before you spend a day on layout. One page that matches an approved offer is a test. Three pages for an offer you cannot run are a folder of unused designs.

How to use it with a CPA offer

Read the offer before you open the builder. Note the countries, the sources, and the action that counts. If you are not approved, stop and request access. Use the preview only to read the advertiser page so your Unbounce headline does not contradict it. Close the preview. Do not copy that address into the button. The address that belongs in the button is the tracking link from the offer, the one that already has your publisher id.

Add sub ids on that link before you paste it into Unbounce. s1 should name the traffic source, so a later report can separate this page from the same offer run somewhere else. s2 should name the variant. If you are not testing yet, s2 can still name the page, so the first version has a label when the second version arrives. Leave room in the other sub ids if a tracker in front of Unbounce also needs to pass a click id. The button URL is allowed to carry s1 through s8. What it is not allowed to be is a naked advertiser URL, a shortened link that strips the query string, or a button that submits a form and then forwards to the preview.

If the visitor arrives on Unbounce with parameters you still need, pass those parameters through to the outgoing button. A builder will sometimes keep incoming parameters on the page URL and then drop them when the visitor clicks out. If a tracker or an ad platform appended a click id, and your button does not carry it onward into a Maxlead sub id, that id dies on the Unbounce page. After you publish, click the button yourself and read the address you land on. You should see your publisher id and the sub ids you intended. Then look in Activity on Maxlead. The click should show the same values. If the click is missing, the button is not the tracking link, no matter what the design file looks like.

Set up the A/B test only after that single path works. Duplicate the variant. Change one thing. A headline, a button label, or the order of the page are reasonable tests. Changing the offer, the country, and the headline at once is not a test of the page. Both variants must use the tracking link. The only planned difference in the query string is the variant label on s2, such as one value for the original and another for the challenger. If both variants share one s2, Maxlead cannot tell them apart, and you will be forced to trust Unbounce’s button-click goal. That goal counts people who left the page. It does not count people the advertiser accepted.

Write both variants so they describe the real action. If the offer is a quote request, the page says quote request. If the offer is an account signup, the page says account signup. Do not use the A/B test as a chance to try a more aggressive claim the advertiser page does not support. The variant that “wins” on button clicks and then dies on rejection was not a better page. It was a different promise. The visitor should be able to tell, before they click, what they are about to do. Hiding the action to raise the Unbounce conversion rate lowers the quality of the click Maxlead records.

You usually do not need Unbounce to receive a postback in order to judge the page, because the variant is already a sub id on Maxlead. If you also want Unbounce, or a tracker behind it, to record the approval, use the global postback or an offer postback on Maxlead. Place {status} and the sub ids you need. Credit only approved. Do not treat a rejected postback as a winning variant. Pending will not fire, so do not mark the Unbounce goal as revenue while the click is still open in Activity.

Publish, then run the country check. A click is stored on the offer only when the visitor’s country is allowed. An Unbounce page targeted at a country the offer does not list will collect ad spend and will not collect Maxlead clicks. Fix the geo on the ad, not the font on the page.

A worked example

A publisher buys native traffic and wants a page for an offer that accepts Canada. The offer allows that native source. The account and the offer are approved. The publisher reads the advertiser page with the preview control, notes the action, and does not copy the preview address.

They build the page in Unbounce. The headline restates the action in the same terms the advertiser uses. The button does not say the visitor will get a prize, a job, or a download the offer does not include. The button URL is the Maxlead tracking link. sub1 is ca-taboola, because the source is Taboola and the country is Canada. The first variant uses sub2 variant-a. They publish, click the button from a test they can identify, and confirm Activity shows ca-taboola and variant-a on a Canadian click.

They duplicate the page for an A/B test. Variant B changes only the headline, to a shorter line that still describes the same action. The button on variant B is the same tracking link with the same publisher id and the same sub1. sub2 is variant-b. They do not send variant B to a different offer to “see what happens.” The native campaign points at the Unbounce experiment URL, not at the advertiser, and not at the preview.

During the test, Unbounce may show that variant B gets more button clicks. The publisher does not scale on that screen. They open the performance report on Maxlead for the same days and read the two sub2 values under ca-taboola. Approved conversions, not button clicks, sit in the conversion count. Earnings are the approved payouts only. If variant B’s extra clicks are mostly rejections, its EPC will be worse even though Unbounce looks happier. The weekly payout will follow the approved rows, not the Unbounce goal. The publisher keeps spending on the sub2 with the better EPC once the click counts are large enough to read, and they leave the other variant in the report so the comparison stays visible.

Mistakes

Putting the preview URL in the button is the mistake that makes the rest of the work irrelevant. Unbounce will still A/B test. The ad will still spend. Maxlead will not see the click, so EPC and CVR do not exist for that traffic, and the weekly payout has nothing to add. Open the live page and click the button before the campaign leaves draft. The address must be your tracking link.

Stripping the query string is the quieter version of the same mistake. Some buttons are set to a pretty domain that redirects and drops s1 and s2. The click might still reach Maxlead if the publisher id survives, but the variant test is gone. Or the whole query is dropped and even the publisher id is gone. After any republish, click the button again. Builders change links when you swap a template.

Testing more than the page is another mistake. If variant B also targets a different country inside the ad account, you are not comparing headlines. If variant B uses a harsher claim, you are comparing promises, and the harsh one can win the click and lose the approval. Keep the offer, the source, the country, and the action fixed. Let s2 be the name of the single change.

Do not use the Unbounce conversion rate as EPC. A page can convert thirty visitors out of a hundred into button clicks and still produce no approved actions. Those thirty are not earnings. Do not edit the winning variant in place and keep the same sub2. The row will describe two designs. Create a new sub2 when the copy changes. Do not send the page live to a country outside the offer. Do not forget that a tracker click id, if you use one in front of the page, has to be mapped into a sub id on the button or it dies at the click.

Do not hide the action to help the A/B test. A button that says “continue” is fine if the page around it has already said what continuing means. A button that implies a free gift the offer does not have is not a design choice. It is a mismatch, and mismatches show up as rejected conversions that never enter the weekly payout.

How to read EPC and CVR before you scale

Turn the winning variant off inside your head until you have read it on Maxlead. Set the date range to the days the A/B test actually ran. Open the performance report for this offer and find sub1 ca-taboola, then the two sub2 values. You are not looking for the variant with the most Unbounce visitors. You are looking for the variant with approved conversions and earnings.

CVR is approved conversions divided by clicks on that row. If variant B has more clicks and fewer approvals, its CVR is worse, and the Unbounce “win” was a loss. If both CVRs look enormous because each row has a tiny number of clicks, you do not have a winner. You have noise. The network is not publishing a minimum click count in this article, and you should not invent one and treat it as a rule of the offer. You should refuse to scale a variant when one extra approval would swing the rate wildly. Wait for a count you would still trust if the next few clicks did not convert.

EPC is approved earnings per click. It moves only when approved payout is added or when clicks are added. Pending actions do not raise it. Rejected actions do not raise it. A variant that produces rejections can show a large click count, a low CVR, and an EPC that does not deserve more budget. Compare EPC across variant-a and variant-b with the same sub1 and the same dates. Scale the higher EPC only when the CVR is not a one-click fluke and the path still checks out: publisher id present, country allowed, action still honest, offer still approved for the account.

Use Activity if the report and Unbounce disagree. A missing click means the button broke. A click with an empty sub2 means the variant label was not on the link. A conversion report row shows the approvals you can relate to the weekly payout. Unbounce’s goal completions do not appear on that payout. When you raise the native budget, raise it toward the URL and the sub2 you just measured. Approved earnings from that decision are paid on the weekly cycle, not when the A/B test hits a confidence badge inside the builder.

Where to start

Unbounce is the page. Maxlead is the network. The A/B test is allowed to change the layout and the wording. It is not allowed to change the button away from the tracking link, and it is not allowed to replace the real offer action with a prettier story. Put the publisher id and the variant on the link, pass through any click id you still need, and read EPC and CVR on the sub ids before you scale.

Get the tracking link from the offer after your account is approved. Start at the Maxlead homepage.