Why traffic changes the Meta Ads setup
Every other objective has a result that proves something happened: a purchase, a form, an install, a person reached. Traffic's default result, the link click, proves only that a finger touched a link. Delivery will find the people and placements that produce those touches most cheaply, including people who tap by accident and surfaces where taps are easy and loads are rare. A traffic campaign can therefore show a superb cost per click (CPCAverage cost of each click.) while the site sees a trickle of visitors who leave at once. The setup moves the definition of success further along, to the page actually loading, and then checks reading outside Meta, because nothing in Ads Manager shows whether anyone read the page. The visit only matters if the page earns it, and the landing pages hub covers what makes a destination hold the people a campaign sends. Whatever the optimisation goal, the visit is decided on the landing pageThe page a visitor arrives on from a campaign, built to turn that visit into one action., so its speed in the in-app browser is part of the campaign setup.
Recommended setup for website traffic
| Setting | Recommended for traffic | Why it differs from other objectives | Feature page |
|---|---|---|---|
| Objective | Traffic | Right only when the visit itself is the goal; outcome objectives exist for sales and leads | Ad formats and placements |
| Optimisation goal | Landing page views | Link clicks need no page load; other objectives already require an event after arrival | Meta pixel setup |
| Placements | Advantage+ placements, then exclusions where the breakdown shows taps without loads | Accidental-click inventory hurts this objective most, since the click is the product | Ad formats and placements |
| Page speed | Mobile page that renders quickly in the in-app browser | A slow page turns paid clicks into lost results directly | Meta pixel setup |
| Exclusions | Recent site visitors excluded from prospecting | Traffic delivery happily returns the same readers; buying them again adds nothing | Exclusions and overlap |
Landing page views, not link clicks
The Traffic objective offers link clicks, landing page views and daily unique reach among its goals. Landing page views count only when the pixel on the destination records the page loading after the click, so delivery has to find people who wait for it. That costs more per result and is worth it: the result is a visit rather than a tap. It also means the pixel must be installed and firing on every destination page, which the Meta pixel setup tutorial covers step by step. Where the site has an engagement event worth optimising for, such as scrolling a recipe or reading an article to the end, the campaign objectives guide suggests adding it.
Placements checked for taps without visits
Meta can show one ad across feeds, Stories, Reels, in-stream video, Messenger and the Audience Network of partner apps. For most objectives letting delivery choose works well, and it is still the right start here. But some inventory produces clicks far more easily than it produces loads, and on a click-priced result that inventory is attractive to delivery. After the first week, break results down by placement and compare link clicks with landing page views for each. Where the gap is wide and persistent, exclude the placement. The ad formats and placements guide explains how manual placement selection works and how to keep the creative cropped correctly for the placements that remain.
Page speed as part of the ad
Meta opens most links in its in-app browser on a phone, often on a mobile connection. Every second the page spends blank loses people who already cost money. Treat the mobile destination as part of the campaign: compress images, defer scripts that block rendering, avoid interstitials that cover the content, and test the page inside the Facebook and Instagram apps rather than in a desktop browser. Core Web VitalsGoogle's user experience metrics. give a shared language for the loading problems, and the speed and mobile guide covers the fixes.
Recent visitors excluded
Delivery does not know that someone read the page yesterday. Left alone, a traffic campaign can send the same loyal readers back again and again, because they click readily. Build a website audience of recent visitors and exclude it from prospecting ad sets, using the rules in the exclusions and overlap guide, so the budget buys new readers. Keep those visitors as a pool for retargeting with a different message.
Engaged sessions as the gate
Ads Manager stops at the page load. Whether anyone read, scrolled or came back is visible only in analytics. Tag every ad with campaign parameters and judge each ad set in GA4Google's event-based analytics platform. on engaged sessions, time on page and the share of visitors who reach a second page. An ad set with cheap landing page views and no engaged sessions is buying arrivals that leave.
A worked traffic setup on Meta
The Brindle Table, an invented recipe website
The Brindle Table earns from display advertising on its recipe pages and wants new readers who return. It spends £40 a day on a Traffic campaign optimised for link clicks across automatic placements.
| Setting | Before | After |
|---|---|---|
| Optimisation goal | Link clicks | Landing page views |
| Placements | Advantage+ placements, never reviewed | Advantage+ placements with one partner-app placement excluded after a breakdown |
| Page speed | Recipe page took about 6 seconds to show content on mobile | About 2 seconds after image compression and deferring the ad scripts |
| Exclusions | None | Site visitors from the last 30 days excluded |
| Quality gate | Clicks in Ads Manager | Engaged sessions in GA4 per ad set |
Over a month at £1,200 the old setup reported 9,600 link clicks at £0.13 each, a CTR of 2.9%, but only 4,700 landing page views, and analytics recorded 2,100 engaged sessions, £0.57 each. In the month after the rebuild the same spend bought 5,200 link clicks at £0.23, 4,600 landing page views and 3,300 engaged sessions, £0.36 each. Returning readers from paid visitors rose from 140 to 410.
The click count fell by almost half and the site gained more readers. Half the old clicks had never loaded a page; the change of goal stopped paying for them, the faster page kept people who would have left during loading, and the exclusion stopped the budget recycling existing readers.
Measuring website traffic in Meta Ads
Ads Manager's results for this objective are Landing page views and Cost per landing page view when optimising for page loads. Keep Link clicks and CTR (link click-through rate) (CTRShare of impressions that result in a click.) in the same view, because the ratio of clicks to landing page views is the clearest sign of accidental taps and slow pages. Outbound clicks count clicks that leave Meta for the destination and are useful where an ad carries several links. Break all of them down by placement before deciding which inventory to keep.
Off the platform, the measure that matters is engaged sessions in analytics, with returning visitors and pages per session behind it. The CPC calculator and CTR calculator make the platform side quick to work out. Check the click to page-view ratio twice in the first week, placements and engaged sessions weekly, and the page speed of the destinations monthly. For a content site, the content strategy framework for awareness helps decide which pages deserve paid readers in the first place.
Where the traffic setup goes wrong
Running Traffic when the business actually wants purchases pays for visitors who were chosen for their willingness to click. Optimising for landing page views without a working pixel leaves delivery with almost no results to learn from. Excluding a placement on a few days of data can remove inventory that was fine. And a campaign judged only inside Ads Manager will look healthy right up to the moment someone opens the analytics.
Common mistakes
- Optimising for link clicks. The cheapest taps are often not visits at all.
- Never opening the placement breakdown. Wasted inventory stays hidden in the average.
- Testing the page on a desktop. People arrive in the in-app browser on a phone.
- Buying the same readers repeatedly. Exclude recent visitors from prospecting.
- Using Traffic for sales. The Sales objective optimises for the outcome directly.
