Best Practices For Targeted Pages

Besides these requirements, which every targeted page must fulfill, there are some best practices every targeted page should aim to achieve. These best practices are based on my own experience, and I have seen them work with many different clients. Some of them may change if the market (Google, the user, the competition) changes so use them as a starting point, then start to establish your own best practices. Unquestioned best practices become worst practices over time!

Best Practice #1: The Targeted Phrase

Define (write down) the targeted phrase of each targeted page. Already very much discussed, every page needs one. There is no single HTML element where you state “this is the targeted phrase” - it is a CONCEPT. The targeting elements listed in this section are in order of priority for their targeting use:

1. Title
2. H1
3. URL
4. Above the fold image filename and alt-text
5. Anchor text

It’s not about stuffing the targeted phrase again and again into these elements - it’s about using reasonable sensemaking variations. If a sensemaking variation of the targeted phrase shows up in the URL, title, H1, the image filename and alt-text you can be pretty sure that Google has got the message and knows what the page is about. The targeting elements are a way to make sure that Google knows what the page is about.

Note: Whenever we talk about a sensemaking variation or a short variation of the targeted phrase then these are all okay:

• partials,
• synonyms,
• different (but correct) spellings (do not target or use misspellings),
• separated and together, and
• singular and plural.

Do not repeat the targeted phrase again and again in the targeting element but, look for, and think about, sensemaking variations. There is a tiny percentage of cases where there aren’t any sensemaking variations, but generally, if you can’t think of variations, then you aren’t thinking hard enough.

It’s also possible that sometimes we refer to the targeted phrase also as the topic of the page. The topic of a page is used in a wider sense as the targeted phrase (for example, if the targeted phrase of a page is [understanding SEO] than the topic is basically “SEO”).

Best Practice #2: Presumed Search Intent

Brainstorm and write down the presumed search intent. If you cannot or will not help the user in some way on their quest, or do not know what the search intent behind a word combination is, then do not even bother to create a targeted page for it.

Remember, you are not just targeting words, you are targeting the user. What does the user want from this page? If your user bounces back from your site in disgust, you will not keep that traffic (or the user), and you hurt your future overall performance as well.

Even though this seems quite clear, it is quite often that I come across non-performing “landing pages” created by “renowned SEO agencies” that do everything right but do not solve the “job to be done” of the user. If these pages do not perform, then this is not because of some esoteric “Google penalty” - it simply means your pages suck.

Best Practice #3: Business Utility

Define and write down the business goal of the page. You must see the value in the traffic that you are targeting. Does it make sense for your business to have specific targeted pages? If you get a lot of traffic, even sustainable traffic, but this traffic has no value for you as it is not your topic and the users will not now, not ever, contribute to your overall business goal, it means these pages do not make sense!

Best Practice #4: HTTPS

Google is now pushing strongly for websites to adopt SSL (a.k.a. https:// with “s” like secure). They announced it as a direct factor for better search traffic performance.

From an SEO, user, and technology point of view, HTTPS is the better solution as it’s the safe bet for the future. You should make sure your http:// URLs redirect (serving a HTTP 301) to your https:// URLs. This can be configured across your whole website. Make sure your website is verified with the https:// URL in Google Search Console. Also, google [HTTPS google] for the specs.

Best Practice #5: URL

The URL is one of the targeting elements. To be seen as a truly search engine optimized URL, it must fulfill the URL rules we determined earlier, including the least important one: URL rule number six (URLs should be a variation of the targeted phrase).

So for example, out of these two URLs:

1. https://www.example.com/content-page/index.aspx?id=0000001, and
2. https://www.example.com/d/understanding-seo,

you can be fairly sure that the user and Google both know the second URL gives more context than the first. URL rule number five (URLs should be short - but not necessarily the shortest possible) is a tradeoff with URL rule six. That means if you are in doubt, use only a sensemaking part of the targeted phrase in the URL. Therefore: https://www.example.com/d/seo-business is better than https://www.example.com/d/seo-is-all-business-activities-with-the-goal-of-getting-found.

If you have a domain name that is part of your targeted phrase, there is no need to repeat it in the URL slug. For example: https://www.books.com/l/fishing-books is not necessary as https://www.books.com/l/fishing is fine. It is better to use a short variation, even if you lose minor context.

Another important aspect of whenever you write a short variation of the targeted phrase into the URL is that the least important URL rule (URLs should include a variation of the targeted phrase) does not violate the most important URL rule (URLs must be unique). So for example, if

• https://www.example.com/a/1239258295/systematic-SEO is equal to https://www.example.com/a/1239258295/SYSTEMATIC-SEO, which is equal, to
• https://www.example.com/a/1239258295/sysTeMATic-SEO, which is equal, to https://www.example.com/a/1239258295/whatever, you have screwed up big time!

Whereby the targeted phrase in the slug doesn’t determine the content of the page. Remember, 1 URL = 1 resource. My recommendation for such cases is to create another DB that matches “systematic-seo” to “1239258295” so basically URLs such as https://www.example.com/a/systematic-seo become possible. Another common way is to make HTTP 301 redirects from

• https://www.example.com/a/1239258295/whatever -> HTTP 301 -> https://www.example.com/a/1239258295/systematic-seo
• https://www.example.com/a/1239258295 -> HTTP 301 -> https://www.example.com/a/1239258295/systematic-seo

But, well, this fix is only halfway - neither is the URL as short as it could be. Additionally, the other URLs are still possible; they just get redirected. At one point you will have lots and lots of internal redirects (which slows down the user on the site).

This is all because someone just wrote some internal link logic wrong, but due to the redirect fix in place, you do not recognize and change this behavior immediately. (Also you communicate more internal not-targeted-pages-but-redirect URLs which slows down your distribution). Don’t catch too much stuff with redirects, if the URL is wrong just let it fail (404) and fix the root cause (the wrong interlinking).

Last but not least, make sure that your URLs fulfill URL rule number two: URLs must be permanent. That means they do not have any dependencies once they are created (i.e. the headline of the page might be used to recommend a URL, but once the URL is created the headline can change a thousand times, and the URL stays the same). All targeting elements must be independently editable! Changing one does not change the others.

Also, don’t use categories in URLs as there is no perfect hierarchy. Categories always change over time, violating URL rule number two.

One last thing (and yes, we SEOs love to talk about URLs): the URL also shows up on the Google Search Result Page. Therefore, a sensemaking URL that aligns with the search query of the user can increase click-through.

My simple URL formula:

For pages where we have complete editorial control and no dependencies: https://www.example.com/%short one or two letter namespace%/%short variation targeted phrase%

• https://www.example.com/d/understanding-seo
• https://www.example.com/l/vienna

For pages where there are dependencies (i.e. user generated content, external database, …) https://www.example.com/%short one or two letter namespace%/%unique id%

• https://www.example.com/p/12345
• https://www.example.com/f/0001xd

Best Practice #6: The Canonical

The canonical element identifies the preferred URL of a web page to Google. The canonical is a very powerful tool. If you make a mistake within your canonical logic, there’s a chance you destroy your traffic. Nevertheless, with just one canonical tag, you can minimize a variety of potential problems.

The canonical is incorporated as a <link> element in the <head> of the HTML of your page, and looks something like this:

<link rel="canonical" href="https://www.example.com/d/understanding-seo">

The canonical gives dominance to the URL the page should have, regardless of what is actually in the address bar of the browser. It is a failsafe way to prevent 80% of everything that can, and will, go wrong with your URL. So, therefore:

• It avoids internal competition with URLs with additional query parameters:
  • URL: https://www.example.com/d/understanding-seo?op=345&source=twitter&foo=bar -> Canonical: <link rel="canonical" href="https://www.example.com/d/understanding-seo">
• It avoids problems with different subdomains gone wild:
  • URL: https://staging.example.com/d/understanding-seo -> Canonical: <link rel="canonical" href="https://www.example.com/d/understanding-seo">
• And, even despite your best efforts, underlying artifacts (such as page IDs), can make it into URLs:
  • URL: https://www.example.com/content-page/index.aspx?id=0000001 -> Canonical: <link rel="canonical" href="https://www.example.com/d/understanding-seo">
  • URL: https://www.completelt-other-domain.com/d/understanding-seo -> Canonical: <link rel="canonical" href="https://www.example.com/d/understanding-seo">

The canonical is page content dependent, not URL dependent. It always shows the URL the content should have, not what it actually has. If Google meets the same page or content via different URLs, Google will summarize all page metrics under the canonical URL. In a way, you can think about the canonical as a redirect that only happens for Google.

Even if there is not a URL variation, the canonical points to itself. The canonical is as static as the content of the page. Make sure that you always include the protocol and the domain in the canonical, too, so that you can fetch (sub)domains gone wild issues (whenever your content is suddenly accessible via two or more different (sub)domains). Quality assurance and checking the canonical after every release is a must.

Known URL variations (for example, because of historical reasons) should always make a HTTP 301 permanent redirect to the correct URL (or just let the wrong/old URL run into an HTTP 404). The canonical is a way to deal with unknown URL variations, a failsave against stuff, that might go wrong with the URLs.

Best Practice #7: Meta Robots

The meta robots tag is a way to provide Google with instructions on how to deal with your page. There are quite a lot of meta robots tags, so please R.T.F.M. We will just deal with the most important one. This is the instruction to allow Google to index your site: <meta name="robots" content="index">, and this is the instruction to tell Google not to index your site: <meta name="robots" content="noindex">.

By default, Google sees every page as set to “index” if there is no meta tag. I recommend that every page (type) has an explicitly set meta robots tag. It must be our decision if something is - from our perspective - worth indexing or not.

Any page that is not a targeted page or the startpage, is deadweight. If the page does not fulfill your QC, set it to “noindex.” It would be much better not to have the page at all (HTTP 404), but sometimes there are reasons to have pages, even though they are not targeted pages.

One good example is “search pages.” It makes sense to have an onsite search engine, and it’s also ideal for the user to be able to copy and share search page URLs such as https://www.example.com/search.php?q=polly+wants+a+cracker. However, you cannot confirm that the search results page fulfills your QC, or doesn’t contain internal duplicate targeting or even internal duplicate content with other prior indexed search pages. Therefore, set these pages to “noindex.”

Note: you can set the meta robots tag in the HTML of your site or via HTTP header. I strongly recommend setting it in the meta robots tag, as this is much easier for regular testing. People just look more often at the HTML code of a site rather than into the HTTP headers.

There is also another meta robots tag called “nofollow.” Never use this. With a <meta name="robots" content="nofollow"> you devalue all links on your page so that Google will not see the links as a vote. And, although rel="nofollow" placed on links makes sense in some edge cases, <meta name="robots" content="nofollow"> applied to a whole page never, ever, makes sense as you are basically crippling Google’s ability to understand your site. Just don’t use it.

Best Practice #8: The Title

The <title> is the most important targeting element. The title gives the whole page content its context. It is visibly displayed in the Google Search results, and it’s the first thing the user sees. Users need to click on the title to read your page.

Note: In this example “Example” is the awsome brand of a tour booking site.

A few rules apply here. The title:

• must be sitewide unique
• must align with the content of the page
• should be about 80 characters long
• should include the targeted phrase, and
• should ideally include the brand of the website

A good format is this:

<title>%targeting phrase% - %brand%</title>

If you are lucky to have a rather short brand (three or four letters) then: <title>%brand%: %targeting phrase%</title> is also a possibility.

If you have enough space (remember, you are limited to about 80 characters), spark up the title by adding elements that might improve click-throughs. For example: “15 Best Hotels in London - Hotels.com” (as people love numbers). Always make sure that your title is the most direct representation of the targeted phrase. If the title states something different than what you actually want to get found for, you are doing it wrong.

Because every targeted phrase must be unique sitewide, so must be every title. That’s why the title is a good way to discover duplicate pages or duplicate targeting on your site.

In Google Search Console go to: Search Appearance > HTML Improvements > Duplicate Title Tags. There you will be able to see pages where there is some title conflict - a sign that something is wrong. Do not freak out about each and every duplicate title, but do solve them if there is a systematic issue on your site that is creating more and more of these duplicate pages.

Best Practice #9: The Meta Description

The meta description is pure marketing as it may show up on the Google Search result page. A good meta description can increase click-through from the SERPs. However, that said, the user will never read this text in its entirety. They will just scan it for a microsecond.

Other than that it has no impact on the page or its performance. It is not part of the content of the page. Rules for the meta description are as follows:

• it must be sitewide unique
• it should be two to four sentences
• it should be descriptive
• it should include a variation of the targeted phrase

I have experimented a lot with meta descriptions. The only thing that I can say for certain is that whenever you include a number in the meta description - one that aligns with the search intent of the user - it gets clicked significantly more often (think telephone numbers, prices, and street numbers).

Sometimes simple unicode bullets like ✓ or similar might also come in handy. If you are too “salesy” in the meta description, Google will not display it to the user. In this case, change the meta description and wait until it shows up. This may take a while.

Best Practice #10: <h1>

The <h1> - the main headline of the page - is the second most important targeting element. It must include a variation of the targeted phrase - and be a good headline! Important: the <h1> must be the biggest textual indicator of what the page is about above the fold.

As a matter of fact, it doesn’t matter if it is <h1>, <h2> or <h3>. In the case that Google renders the page, it picks the biggest headline, and uses it as the main headline, no matter if it is <h1>, <h2> or <div>. But, as Google does not always render the page, just make sure that <h1> is the biggest headline, and that the headline clearly states what the page is about.

A clear headline is not only an official Google requirement, but it’s also about user experience. If the user types in [staten island ferry], sees a search result with a big blue underlined link to “Staten Island Ferry - Example.com,” they will most likely click on it. If the user then lands on a page which does not clearly state that it is about Staten Island Ferry, the user will bounce back.

Pages with big, descriptive headlines above the fold that actually tell you what the page is about work better for the user, and therefore perform better in Google.

It’s best practice not to have the title and the <h1> as the exact same text. On the other hand, they should not be completely different as they are both based on the targeted phrase. Using the example of the [staten island ferry]:

• Title: Staten Island Ferry - Example.com
• H1: Staten Island Ferry Ticket Prices and Timetable
• Incorrect H1: Get the next boat!

The <h1> does not have a length restriction, other than the ones that usability sets (a paragraph of text is not a good headline). There should only be one <h1> on a page (as an element and the way it is displayed). Therefore, there should never be a shadow of a doubt what the main headline of the page is.

Best Practice #11: Above the fold image filename and alt-text

The image filename and the image alt-text of the above the fold, main image is the fourth most important targeting element. We are talking here about big, main images that are an essential part of your main content. We are not talking about thumbnails. (The main thing you should worry about when using thumbnails is to make the ones above the fold load blazingly fast, and, if they are linked, have an alt-text that fits the linked pages.)

If, and only if, it makes sense for your page to have an image (and most of the time it does - people love images) and especially if the image has a lot of web page real estate above the fold (most of the time it does), then your image should have the following requirements (we’ll take them step by step):

• The image should fit the topic of your page: Google knows what your image displays quite well.

  The advancements in image recognition by Google are very, very impressive. So displaying something completely off topic can hurt your performance. Never display any porn or erotic images, unless of course, you are targeting adult queries.

• The image should be unique: The main image of a site is usually above the fold and takes up the most real estate, so it should be unique.

  If you do not have unique images it’s a lost opportunity, but not a major problem. So, what exactly is unique? You should be the first one to communicate this image online. So no stock photos, which is a good thing, because stock photos suck.

  This usually means you have to take your own photos. Which is easy - just use your mobile phone’s camera. From an SEO point of view, and from a user’s point of view, these images are good enough in most cases. And, they are much better than the 1,000 stock photos of [happy woman eating salad] (google it).

  Be creative: in a half day photo shoot, you usually can create one hundred unique, entertaining pics - all of them great unique content. On the other hand, if you use stock photos, you have wasted most of your above the fold real estate (what most users see, and therefore has the most value for Google).

• The image URL must fulfill the URL rules. There is only one small thing that differs between targeted page URLs and image URLs: the fact that you can be more relaxed about the “short” part. It’s also fine for properties such as “big,” “small,” “300x500” to be used in image URLs.

• The image must have a descriptive alt-text that includes a variation of the targeted phrase: The alt-text gives additional context to the image, especially if the image is linked (as this one should be, see next point).

  Here, it’s best practice to use approximately five very short words that are at the same time descriptive, as well as a sensemaking variation of the targeted phrase. If in doubt go for more descriptive instead of using a non-sensemaking but targeted phrase (if you do image optimization right, these two properties aren’t exclusive).

• The image should link to a bigger version of itself: Images are always indexed in the context of the embedding page.

  Google Image Search prefers larger images (at least one side of the image should be 1200px). If you want (and you should want) to be displayed in Google Image Search and Image Onebox (the Google Images Search embedded results in the normal Google Search), then you must link from your embedded version of the image to the bigger version.

  The image must be more or less the same content. Slight variations are okay (a small copyright notice, or showing a partial in the small images and the slightly fuller version in big), but showing two completely different images will not work (Google is not stupid). In the end, an embedded image markup should in its simplest form look something like this:

<a href="https://images.example.com/i/understanding-SEO-big.jpg">
<img src="https://images.example.com/i/understanding-SEO.jpg"
alt="Book Cover Understanding SEO" />
</a>

Note: It’s okay to have images on another subdomain. It’s best practice to have the smallest images (kb wise) possible, so please google and use [image minifiers].

Best Practice #12: Diversity of Content

Users don’t read your content. They really don’t. During my workshops, I always ask who can remember the last article they read in full, from top to bottom. Very few people can. Then I ask if they can remember the last article they googled and read in full. In most workshops, no one could, in every other workshop one person could. Even then it was always a GO-query (the user knew of the article and wanted to reach exactly this article).

Users coming from generic know- and do- queries don’t read your content! They scan it. They have a “job to be done” in mind and want it to solve in the most efficient, fastest, laziest way. So how do they consume our content? They quickly scan or skim it, looking for a fragment of information that will let them advance to whatever their goal is.

• Some of the users like short paragraphs of text,
• some like a bullet point list,
• some like infographics,
• some like a video,
• some like infoboxes, and
• some like structured instructions.

Others just want to have a list of things to click. There is not one type of user: we are all users, we all like to do things our way. With SEO, you have to cater to all types of users.

Google knows that all users are different. That’s why Google strives for diversity on the first SERP by showing different kinds of pages, that include different kinds of content, like:

• an article (paragraphs of text)
• addresses
• bullet points
• tables
• images
• videos
• youtube videos
• structured content like recipes and opening hours
• infoboxes, and
• maps

A simple way to achieve visibility on the first page of the Google SERPs, is to look at what isn’t there. If there is no page with a video, then it makes sense to create a page with a video. If there are no results with tabular data, create a targeted page with an extensive table, and so on. You will show up (well, you will show up, if your site is fast too)! Your brand and user experience will determine if you stay there.

In general, you should prepare your content on your targeted pages in as many “sensemaking” ways as possible. This can and should be combined with the SEO process of “How do we update x% of our targeted pages?” which will be discussed later in this book. Don’t try to do it all at once, but over time.

What content do I mean when we talk about “diversity of content?”

1. Paragraphs of Text

One or more short paragraphs of text are best structured in a way to make it more easily readable with tags such as:

• <i>,
• <b>,
• <strong>,
• <small>,
• <sub>,
• <sup>,
• <del>.

It’s not about the value of these tags; it’s all about readability. Keep in mind you must strive for content diversity, not content quantity. If you have a long article (four or five paragraphs of text), this page will not perform significantly better if you just add another paragraph.

Important note: Do not push the targeted phrase into the paragraph of text if it does not fit in there. The targeted phrase is necessary for the targeting elements. If it fits naturally into the text - most of the time it does anyway - then there is no issue, but forcing it in there has no positive upside.

If an SEO comes your way and tells you to write “SEO texts with relevant keywords in the first paragraph” just get rid of them - they are an idiot. The targeted phrase should just be minded in the targeting elements. All other elements must “just” be on topic and fulfill the user’s intent. Stop thinking of the targeted phrase as “keywords,” think user search intent.

2. Bullet Points

If you want to list something within an article or body of text, use bullet points <ul> <li>, or an ordered list <ol> <li>. These simple lists offer a better user experience and more value than just another paragraph of text.

But there are also other kinds of lists. A list of short teasers, either in the form of:

• headlines,
• (half)sentences,
• structured information (for example, prices), and
• thumbnails for example.

Each list item is on its own non unique-internal or external duplicate content. We call targeted pages where most of the main content is a unique combination of content list “targeted list pages,” or TLP.

Whenever you are creating lots and lots of targeted list pages, you must always be sure that the lists are unique. Each list (independent of ordering) must only exist - internally and externally - once, otherwise it must be seen as internal or external duplicate content.

On a TLP, it’s important not to push the list down below the fold. If the user most likely wants a list, give them a list, not an article. The unique combination of content usually links to other targeted pages. Make sure that the whole block is not linked, but certain elements separately (thumbnail, headline) so that you deliver sensemaking internal anchor text. Additionally, a list must really be a list. Each list item must be part of an unordered list <ul><li> or ordered list <ol><li>.

A typical “rookie mistake” is to disfigure completely valid targeted list pages with nonsensemaking, general (crap) text (a.k.a SEO text). Maybe a short introduction text before a list might make sense, but most of the time it doesn’t so don’t do it.

3. Tables

If you compare something on your page, or you have any other kind of content that would benefit from a tabular display of the data, use a HTML table.

4. Embeddable Content (Videos, Maps, and More)

Most embedded content is external duplicate content. Pages with embedded content only work if they are placed within the bigger context of other content, making it part of a unique combination of content.

If it makes sense for the user experience to embed some relevant content (for instance, a Youtube video, a map, a comments box-widget, or a photo gallery) on the page, then simply do it. Most of the time (and definitely for videos and maps) Google knows that this page has a video or something else embedded. Therefore, these elements add to the diversity of content.

In other cases where Google does not know, can’t crawl, or will not crawl the embedded content, it is still worth it if it has a positive impact on the user experience. But remember always to use asynchronous, non-unblocking embed codes methods. Third party content must not make the page any slower. The speed goals of a page are always absolute, and excuses like “third party crap makes our site slower” do not excuse anything and are not acceptable.

One note about galleries: only the first image that is shown when the site was loaded can be considered the main image of the page. All other images are seen as hidden after user interaction, even if it’s a carousel. (Note: Don’t use carousels. They are a horrible, horrible user experience.)

5. H2 and Other Headlines

If you have different sections within your main content, label them with different subheadings. Some of these subheadings might include a sensemaking, short variation alternative of the targeting phrase. Do not artificially lengthen all subheadings with the targeted phrase again and again.

6. Diverse Images

The big, above or near the fold main image is the most important image of the site page. However, having different kinds of images on a page, for example, a photo, an infographic, a diagram, or an illustration, contribute to the diversity of content. Similar to the main image they should have optimized URLs, alt-text, and link to a bigger version of itself. Whenever embedding images make sure the speed goals of the website are still met.

7. Structured Content

Structured content usually comes in the form of bullet points, definition lists (infoboxes), and key value pairs. Any content than can be sensemakingly structured should be structured using proper HTML (<ul><li><ol><li>, <dl><dt>), i.e.:

Best Practice #13: List of links to other targeted pages

Going back to my list of best practices, every targeted page should link to other targeted pages and be linked from at least one other targeted page. Therefore, every targeted page should have a page specific list of links to other pages on the website.

Sometimes this list is part of the main content, but quite often it’s part of the supplemental content (for example, on the right-hand side, below the main content). It can be five links; it can be ten links, or it can be 50 links. It can be one list, or it can be different sublists.

As always, these links should make sense. That said, add some “serendipity” to each page, and the requirement that every targeted page has a link to another targeted page will be fulfilled (note: serendipity is just a fancier word for randomness). You can give these link sections the headline, for example:

• “Serendipity”
• “From the Archive”
• “Other Stuff”
• “What other people viewed”

If you have a site with five million targeted pages, and you add only three serendipity links to each of these pages (and within these, there are links and the page does not link itself) you have a 95% probability that each and every page has at least one link from another targeted page.

I strongly recommend creating sensemaking and serendipity link sections on your site. However, be aware that you don’t fall towards the other extreme and only link serendipitously on your site. More important targeted pages should have more links than less important targeted pages. And make sure your list of links is an actual HTML list.

Best Practice #14: Internal Anchor Text

The internal anchor text is a targeting element. Making sure your internal anchor text is sensemaking, providing an indication of where the link will go, if clicked. Make sure your internal anchor text is a short or long sensemaking variation of the targeted phrase or at least describes the topic of the page. Or to put it more simply: never use phrases such as “click here.”

• DON’T DO THIS: <a href="/a/understanding-seo">CLICK HERE</a>, but
• DO THIS: <a href="/a/understanding-seo">Understanding SEO</a>, or
• DO THIS: <a href="/a/understanding-seo">SEO</a>, or
• DO THIS: <a href="/a/understanding-seo">new SEO book</a>.

Make sure all links are immediately recognizable as links (the trick is to underline them - blue and underlined is awesome)! This is all very good for your usage values as well. Users love to click links. They like to click on links which align with their overall intent, and they will only click on a link if they know it’s a link.

Best Practice #15: Code Quality

Google is great at reading crap HTML markup, so it’s not necessary that your markup strictly fulfills the W3C HTML Specification and throws no errors with the W3C Validator.

But your markup must:

• be damn fast
• be used as HTML was intended to be used
• render in a sensemaking way for common browsers

Use headlines as headlines, use lists as lists, use tables for tables, use <ul><li> for bullet point lists, normal links as links, and so on. If an experienced developer cannot make sense of your HTML markup, then there is a good chance Google can’t either. Beware of:

• <div>- soups (markup where everything is a div),
• headlines used for text, and
• all websites which were created by WYSIWYG editors from 2000’s.

And did I mention it must be damn fast? Best practice is to have nothing in the HTML file that does not get initially rendered in the browser. If you have 10,000 lines of HTML boilerplate (for example, navigation with dropdowns) and five lines of actual content, do not wonder why your site does not perform in Google.

Best Practice #16: Schema.org

Schema.org is a technical specification to enrich your page with structured content so that machines (mostly Google, who are one of the sponsors) can understand it better.

All Schema.org specs can be found at http://schema.org/docs/schemas.html. But wait! Do not implement them right away. Google only uses a subset of these (google [Rich Snippets] and [Structured Data Google I]). As with all technical specifications, the first thing you should do is R.T.F.M.

As you will see in the specs, Google documentation support allow structured markup for:

• products,
• recipes,
• reviews,
• events,
• software apps,
• videos,
• articles,
• logos,
• corporate contacts, and
• site links search box (and this list is constantly expanding).

Thankfully, the people Schema.org was gladly deprecated in 2014, so you don’t have to worry about that anymore. In the cases of products, recipes, reviews, events, software apps, videos, and articles, you get better visibility in the Google Search Results with so-called “rich snippets” and “rich cards”. So for example, a recipe rich snippet would look something like this:

And, an event rich snippet would look something like this:

As you can see, this offers a clear advantage in search results. For the other types of structured material that do not have rich snippets, Google does something with it internally, and it contributes to your diversity of content.

My recommendation is always to implement Schema.org for all schemas where Google has announced official support. Apply it as described in Google documentation and ignore all the rest. If you are in a competitive vertical where Schema.org use is common, you must use schema.org.

As always, use schema.org in a sensemaking way. Lots of websites spam Google (and themselves) with rich snippet markup that leads to rich snippets, but was actually intended for other uses. This is called “rich snippet spam.”

Quite common is the star rating, which was intended for review sites of any kind, but gets used by showing up on pages without visible reviews. Google catches up on this sort of spam (most of the time, the site just loses the rich snippets, and nothing else happens).

However, misleading rich snippets lead to a bad user experience and a lower click-through rate in the long run. So, yes, you can mark up your non-recipe-article to look like a recipe, and get a beautiful image in the SERPs, but then the users will think you have a recipe, and just bounce back after they have visited you and didn’t see the recipe. So just don’t do it.

There are so many other things you can do SEO wise, and making rich snippet spam is a very bad investment of your time and resources.

Note: as Schema.org is invisible meta data with no visible feedback it it breaks, please always (after every release) test your implementation with Googles [Structured Data Testing Tool] and regularly check “Google Search Console > Search Appearance > Structured Data” and “Google Search Console > Search Appearance > Rich Cards” for all so common “everything is broken again” errors.