With global ad spend reaching $332 billion in 2020, it’s safe to say that ads are the fuel of online publishing.
For some publishers, paywall content is the leading source of revenue, but the majority still rely on ads as their main revenue source. All this should come as no surprise, considering ads have been an integral part of newspaper publishing since 1704!
The SEO community has had its ups and downs when it comes to ads. Probably more downs than ups. In the past decade, Google announced a number of algorithm updates concerning ad usage. These announcements definitely created waves in the community. SEOs managers who were accustomed to solving algorithm update puzzles on their own, were getting detailed explanations on the upcoming updates. This led to a climate perfect for breeding rumors and myths around the subject of ads and SEO.
To better understand the relationship between the ads and SEO, we first need to examine the setting. On the one hand, it’s in Google’s interest to serve users the content they clicked on from search results as seamlessly as possible. Publishers on the other hand aim to increase revenue by showing ads. This conflict of interest leads Google to try and reduce the number of ads that obstruct and distract users from consuming content. The main logic behind these algorithm updates is to simply enable a better user experience, especially in a mobile-first online world.
Fun Fact: Back in 2003, SEOs believed that hosting AdSense on your site might improve your rankings.
The impact of ads on the SEO debate can be traced back to 2012, when Google announced the page layout algorithm update, later known among SEOs as “Ads Heavy”. Here is what the update was about, in essence:
“If you click on a website and the part of the website you see first either doesn’t have a lot of visible content above-the-fold or dedicates a large fraction of the site’s initial screen real estate to ads, that’s not a very good user experience. Such sites may not rank as highly going forward.”Matt Cutts, 2012
The update should have impacted less than 1% of searches globally. For comparison, Google estimated that passage indexing will impact nearly 7% of searches globally. The page layout update ended up impacting 0.7% of sites when it was first rolled out. It has since then hit twice more, in 2012 and 2014, until it became an integral part of the algorithm.
Why do we bother digging up an algorithm update almost 10 years old now? Because it is still relevant and extremely important today for sites running ads. Back when it first rolled out, it caused a lot of turmoil, and it’s still a source of misinformation and misunderstanding regarding the impact of ads above-the-fold. To cut a long story short, having ads above the fold is fine, as long as it is not done to an excessive degree.
“This algorithmic change does not affect sites who place ads above-the-fold to a normal degree, but affects sites that go much further to load the top of the page with ads to an excessive degree or that make it hard to find the actual original content on the page.”Matt Cutts
What is a “normal” degree?
If you want a more up-to-date take from Google regarding this issue, this is how John Mueller explains it:
How the Content of Ads Impacts SEO?
It seems that some publishers implement ads on their site with a “set it and forget it” mindset. This is especially true, but not limited to, Google AdSense. Publishers do it this way because they believe the ad network filters out problematic ads for them. And it’s true to some extent – some ad networks do filter out content that might be offensive or harmful to users.
The thing is, some ad networks are pickier than others. Each network has its guidelines regarding what ads are or are not acceptable. In addition, there are regional regulations that govern online advertising. And on top of that, most ad networks rely on humans to evaluate the submitted ad content, and humans are subjective creatures – what’s considered offensive by one person is completely okay with another.
It’s important to note that in terms of search engines, even though some ad networks do not enable you to directly control the content of the ads they show on your site, you are held responsible for the overall quality of the ads displayed. This is an official Google policy as reflected in section 2.4.3 in the search evaluator’s guidelines.
How does Google evaluate the quality of an ad?
There are two ways Google can evaluate your hosted ads quality: using a human evaluator or by applying a technical solution.
How Googlebot understands ad content?
Unfortunately, no publically official resource explains how Googlebot interacts and interprets ads content. However, two tools could help you understand how your ads quality is perceived:
- Ads experience report: The ad experience report is one of Google search consoles legacy reports. If you have any ad quality issues, they should appear here. However, trusting this report as a single source of information is not advisable. The report was recently added to the page experience report on GSC. So we could expect it to be upgraded and fit into the not-so-new GSC sometime in the near future.
- Chrome built-in Ad filtering: Chrome has a built-in adblocker, which should filter out intrusive or misleading ads. If you want to check whether it blocks some of your ads (only on desktop), you can click the lock icon left of the address bar, and then choose “Site settings”.
Change the drop down to allow:
You can then compare versions of your page with ads blocked and ads allowed, and see if you find any differences. Chrome is supposed to pop up a dialog box when it blocks some ads. But in order to be sure, it’s better to conduct the above check.
Can Googlebot (or other bots) understand the content of ads?
As machine learning models get stronger and smarter every day, it is 99.9% safe to assume that YES, bots “understand” the content of ads very well. Without going into too much detail and with the intent to keep this article focused on SEO, let’s look at two machine learning models used by Google algorithms that can help us understand how bots “understand” content.
- Natural Language Processing (NLP) using BERT (and SMITH): Simply put, BERT is a machine learning model that helps Google better understand the intent behind search queries. SMITH does the same, but it can be used for long-form documents (BERT is limited to short texts). It’s important to note that BERT is not used for analyzing ad text, but you can only assume Google uses some NLP Machine learning model to do just that. The closest I could find is this ads sentiment analysis API.
- Image Recognition: To understand how banners or image ads are interpreted, you can use Google’s vision API. If you see ads on your website that might hurt your SEO, this would be a good way to find out what a bot would “see” in the image.
Does Googlebot crawl ads?
There is no unequivocal answer to this question. If you want the most informed answer, you should consult the ad network, and ask them to check their logs for Googlebot visits. But sometimes they won’t have a concrete answer either. Instead, we can try and simulate a bot’s behavior on a page and by doing so, indirectly learn how it behaves around ads.
- Open a page in Chrome incognito tab (make sure your Chrome is updated).
- Open Chrome DevTools.
- Click [Esc] to open the drawer and choose the “Network conditions” tab (if you can’t find it click on the dotted menu).
- Untick the checkbox next to “User-agent” and choose Googlebot or Googlebot Smartphone for mobile from the dropdown.
*If you check the mobile version, make sure the mobile emulator in the upper left corner is switched on.
- Refresh the page.
You can now have a look at how Googlebot sees your page. First, check to see if any of the ad placements are gone from the page. This could indicate that the ad network is blocking Googlebot, usually with a robots.txt or other method.
Another easy way to see which ad network resources are blocked by robots.txt is the Mobile-friendly test tool, or GSCs URL checker.
As we can see, Googlebot does not render the ad itself, so it is safe to assume that it also doesn’t crawl it. However Google does use AdsBot to crawl ads, and here is the thing: your crawl budget is shared between organic and ads. This might imply that your organic crawls could be negatively impacted by AdsBot crawls.
GSC URL checker
Enter a URL to the tool.
Click on: View crawled page.
Click More info > Page resources
Mobile friendly tool (if you can’t access GSC)
After you enter a URL to be checked, click on “Page loading issue”.
You will then be able to see the blocked resources.
Human evaluators & manual penalties
In contrast to technology-based evaluation of ad content, where we can only infer how it’s done, we know for certain what Google’s human evaluators are instructed to do. Here is what Google search quality evaluator guidelines have to say about it:
“We expect Ads and SC (supplemental content) to be visible. However, some Ads, SC, or interstitial pages (i.e., pages displayed before or after the content you are expecting) make it difficult to use the MC (main content). Pages with Ads, SC, or other features that distract from or interrupt the use of the MC should be given a Low rating.”
Content of ads is considered distracting when it falls under one of the following categories:
- Sexually suggestive images
- Grotesque images
- Porn ads on non-porn pages
- Titles or images of the ads are shocking or disturbing
If you require a visual example, please visit the guidelines.
The Ratio of Ads to Content (Ads density)
Pick a number between 0-100. If you have been in publishing long enough, you should be thinking about hmm, thirty?! Did I get this right? 30% – that’s the “magic number”. This number comes from research conducted by the Coalition for Better Ads. But what does it mean?
30% represents the density of ads on mobile as a percentage of the vertical height of the page. That is at least the official definition. But here is the thing: while the Coalition for Better Ads takes into account only the main content of a page (excluding headers, footers, and site navigation), Google defines it as 30% of the initial viewport. This could be an honest mistake, but we need to consider that this is the number used in the Lighthouse beta ads inspection tool. So for SEO purposes, I believe 30% of the initial viewport is more relevant.
In relation to SEO, it’s highly important to find out how humans interact with your page and not only how Googlebot interacts with it. 30% is a nice number to keep in mind, but you can have a very destructive experience also with lower ad density. It all boils down to the type of ads you serve and their location on the page. To reduce the impact of your ad-density on UX, it’s important to conduct ad balance experiments.
How Infinite Scrolling impacts ad density?
As mentioned earlier, Google recommends keeping ad density lower than 30% of the initial viewport on mobile. But what if you have an infinite scroll widget?
Google uses something that is called “frame expansion”. They first try to render the page with a ±10,000px height viewport, and then try to scroll the page down once. Your infinite scrolling feed should technically not expand more than once, meaning it should only show two chunks of it in the viewport when scrolled down once. This should definitely not come close to 30% of a 10,000px viewport. If your feed does not behave this way, we recommend you ask the feed provider to fix the implementation.
Intrusive Interstitials and Popups
Back in 2016, Google released the following announcement:
“Pages that show intrusive interstitials provide a poorer experience to users than other pages where content is immediately accessible. This can be problematic on mobile devices where screens are often smaller. To improve the mobile search experience, after January 10, 2017, pages where content is not easily accessible to a user on the transition from the mobile search results may not rank as high.“
Many in the SEO community freak out when they encounter such a direct announcement from Google. They tend to recommend avoiding interstitials at all costs. But this is far from the ideal reaction…
SERP traffic vs. in-site navigation
When a user clicks through search results to a page, Google wants that user to be able to find the content they searched for easily. When it comes to navigation within the site, there is no problem showing users interstitials. This is also true for direct traffic! See below what John Mueller has to say about it.
Desktop vs. Mobile
When looking into Google’s announcement, you’ll notice that there is no reference to desktop. Does this mean that it’s 100% safe to use interstitials on desktop? Probably not. We also asked John Mueller on the subject (unfortunately it was after the session recording ended), he replied that intrusiveness refers to desktop and mobile alike. However interstitials on desktop need not be as content blocing
GDPR, CCPA, and cookie notices
Google tries to identify this type of interstitials, and they are not considered intrusive.
Interstitials dos and don’ts
If you want to use interstitials in a way that respects the user experience, it’s best to follow these suggestions, derived from Google’s own- Traffic web interstitial solution (currently in beta):
- Automatically pre-load and render before navigating away from the page without delay when the user clicks a link on the page.
- Have a clear exit option.
- Legal Interstitials – Cookies or age verification.
- Login dialogs for private or gated content including paywall content.
- Banners that use a reasonable amount of screen space and are easily dismissible.
- Show a popup that covers the main content upon navigation to a page from SERP, or while engaging with the page.
- Display a standalone interstitial that the user has to dismiss before accessing the main content.
- Use a layout where the above-the-fold portion of the page appears similar to a standalone interstitial, but the original content has been inlined underneath the fold.
- There are some more best practices regarding Admob interstitial implementation on apps. Those could be beneficial when trying to understand Google’s interstitials mindset.
Exit Intent popups
Not all pop-ups are created equal. Take a closer look at the following directive from Google:
“Showing a popup that covers the main content, either immediately after the user navigates to a page from the search results, or while they are looking through the page.”
If you read between the lines, one question comes to mind. What if the user has finished consuming the content, is it okay to show an interstitial than? According to the Coalition for Better Ads, you can show exit intent banners in four scenarios:
- When the user starts to leave the page (without interfering with the user’s departure)
- The user has been inactive for more than 30 seconds on a page that does not contain a video
- Once a user has reached the end of the first article on a page
- If a user purposely navigates to another tab and then returns to the open page.
How to measure the impact of interstitials on SEO?
As John Mueller explained in the above video from Google webmaster’s office hours, there is no way to measure the impact of interstitials on SEO. There is also no way to know if your site was impacted by showing interstitials.
Labeling Ads & Differentiating Them from the Main Content
This one is more of a guideline than an SEO rule. But we think it is extremely important because it receives a lot of attention on Google’s QRG.
Your ads must be labeled as ads. You can label them as “ads,” “sponsored links,” “sponsored listings,” “sponsored results,”. It doesn’t matter how you choose to label, just that you label them.
Here are two examples of deceptive labeling that would result in lower ratings:
- Content on the right is labeled “Top Posts & Pages.” It is unclear whether these are supplemental content or ads.
- Ads and links to other questions (misleadingly labeled as “Relevant answers”) displayed prominently, which users may mistake for answers to the question. It takes a moment to notice that this page has no answer.
Ads Impact on Performance (Page load speed)
At the time of this writing, Google has not yet rolled out Core Web Vitals (CWV). So if you are reading this in the future, you might as well just roll your eyes now. Core Web Vitals was introduced by Google in May 2020. Later that year, Google announced CWV would become a ranking signal as of May 2021. This announcement has triggered a chain reaction in the SEO world equivalent to an oversized tsunami.
To put things into perspective (which is much needed in this case), CWV will not be an SEO armageddon, as Google officials recently expressed. Check out these quotes from John Muller, Martin Splitt, and Danny Sullivan.
After sounding an all-clear signal, let’s get down to business. Ads can be a highly demanding resource especially when implemented recklessly. The main issue with ads is that they are implemented by third parties, so we do not have much control over the implementation itself. The impact ads have on page-load should be a subject of its own blog post (stay tuned), but I will try to go over the main points here, just to give an initial idea.
How ads impact performance?
Performance has a very broad definition. There are a lot of metrics you can track, and even more tools you can measure “performance” with. To keep it simple, we should limit the discussion to the CWV metrics and measurement tools provided by Google (page speed insights, Lighthouse, GSC, etc.). We also need to distinguish between Lab data and Field data.
CLS – Cumulative layout shift
CLS is the metric that ads potentially impact the most. Put simply, a layout shift is when content pushes down other content on the page when it is loaded. The way to tackle CLS is to statically reserve space for the ad slots.
LCP – Largest contentful paint
LCP measures the time it takes to load the largest element within the viewport. Ads generally should not impact LCP directly, but they could have an indirect impact. Ad networks inject ads to pages using a JS (script). The script aims to load as fast as possible, but it usually loads asynchronously.
Loading the script with an ‘async’ attribute means that the script will not be render-blocking, but the script still needs to be loaded, and it loads without waiting for other components, like your hero image. So it could potentially impact the time it takes to load your LCP.
As you can’t ‘defer’ ad network scripts, the solution here needs to come from the ad network, which should make sure they are loading the script in the most efficient way possible.
From the publisher side, you should make sure your script is loading with the ‘async’ attribute. Also, preloading the script could come helpful in some instances.
FID – First input delay
FID measures the time between the user’s first interaction with the page (i.e. clicks a link, a button etc.) to the time when the browser can begin processing event handlers in response. FID can only be measured on the field. So you can’t optimize for it. The metric that correlates best with FID in the lab is TBT: total blocking time.
Without going into too much detail, TBT measures the total time the main thread was blocked between FCP and TTI. In the context of ads, there is not a lot you can do here besides what was mentioned earlier regarding LCP (see above).
How to measure the impact of ads on core web vitals?
Unfortunately you can’t isolate the impact of ads on CWV from other components on your page. The closest way to measure the impact is by blocking the ad network requests using Chrome DevTools and comparing your Lighthouse scores with blocking vs. without blocking. Remember to conduct this test on an incognito window, without any Chrome extensions running in the background.
Step 1: Open a page you would like to examine, run Lighthouse, and record the results you are interested in. We suggest focusing on CLS, LCP, and TBT on mobile.
Step 2: Close the page and reopen it in a new incognito window. Open DevTools again and navigate to the Network tab. You will need to block the request the ad network makes to measure the impact of ad placement. You can filter down by the ad network name to find out the requests it makes. When you find the request whose impact you want to measure, right-click on it and choose “Block request URL” or “Block request domain” to block all placements from this ad network.
Step 3: Run Lighthouse again, and compare the results you get with those you got before the blocking. The difference represents the impact a placement or ad network has on your page CWV. Keep in mind this is not accurate, but it’s close enough to give you an indicative number.
Best practices for improving ad performance
If you talk with technical SEO specialists who are familiar with boosting page speed, you might be advised to lazy load your ads. This approach should be handled with extra care! Lazy loading ads will reduce ad revenue significantly. The question you need to ask yourself is how much extra organic traffic you will gain from this move? Will it cover the losses from ad revenue? I will again quote John Mueller on this:
“It’s useful to keep that in mind when it comes to Core Web Vitals. It is something that users notice. It is something that we will start using for ranking. But it’s not going to change everything completely.
So it’s not going to… destroy your site and remove it from the index if you have it wrong. It’s not going to catapult you from page ten to number one position if you get it right.”John Mueller
If you do want to make significant improvements to your ad performance, make sure your ad network complies with the following:
- Serve images in next-gen formats (mainly WebP)
- Serve images from a well known CDN
- Lazy loading ad images as opposed to the ads themselves.
Keeping track of CWV – Field Data
It’s currently not possible to track down the impact ads have on your CWV field data. The data you receive through GSC, CRUX or GA can’t be broken down into measuring the partial contribution of ads to your overall CWV score. We hope Google will make this data available in the future.
Fun Fact: If you think ads make your site heavy, you should keep in mind that the first internet banner ad (468x60px) for AT&T was downloaded using a 28.8 kbit dial-up connection 🙂
TL;DR: How to Minimize Ad Impact on Your SEO
To eliminate the impact of ads on your SEO, follow these steps:
- You can have ads above-the-fold to a reasonable degree (make sure that the searched content is available for the user).
- You are responsible for the content of ads on your website even when they are coming from third parties.
- You should monitor the quality of ads as much as possible.
- Keep ads density under 30% of the viewport, and conduct tests to find out how they are best located.
- Interstitials are not taboo, you can use them without impacting SEO.
- Ads should be labeled properly and be easily distinguishable.
- Ads, like any other element on a page, influence page load. When implemented correctly, the toll is reduced to a minimum.
- Lazy loading ads impact revenue negatively. Lazy loading ads images and optimizing them will do the trick, without compromising revenue.
- When in doubt, we suggest going through Adsense ad placement policies to get the “spirit of the commander” regarding the dos and don’ts.
Ads are an integral part of online publishing, and that’s why Google announcements about the use of ads and interstitials in the past decade have triggered unbalanced reactions among the SEO community. Having ads and interstitials on your site can go hand in hand with SEO, and when implemented correctly, shouldn’t have ANY negative impact on your organic visibility.