Search Engine Optimization Best Practices

Search engine optimization (SEO) helps make your web pages easier to find through search engines like Google. An SEO-friendly website is well-structured, readable, and accessible, ensuring users can quickly find relevant information.

Many factors influence search rankings, but there are key best practices that improve visibility. Using clear titles, descriptive system names, and relevant keywords helps search engines categorize content effectively. Proper content structure also enhances how pages are indexed and displayed in search results.

The following guidelines outline essential SEO strategies to create a web experience that is user-friendly, accessible and search engine-friendly.

Introductory SEO Questions

What is SEO?

SEO stands for search engine optimization, and it is the process of optimizing web experiences to better serve both users and search engines. The goal of SEO is to achieve more organic (non-paid) website traffic through web pages that are deemed higher quality and thus ranked higher by search engines.

Why should I aim to improve my web pages’ SEO?

How do Search Engine’s rank content?

Search engines use thousands of signals to rank pages for queries, but there’s a few that are typically agreed upon to prioritize.

Here’s Google’s documentation on how search works.

Does every page need to be SEO-optimized?

Sort of, but not really. Let’s explain further.

All pages, regardless of intent, should be technically optimized for best performance.

After a page meets the technical standards for SEO and accessibility, then we must consider intent.

Pages for our internal audience (current faculty, staff, or students) will likely not need to have the same optimization that a page that is trying to reach prospective students is.

How often should I update content?

Main pages should be updated at least twice per year, but the goal for heavily trafficked pages should be even more often.

Content audits should be done annually at the minimum, and preferably twice per year.

How do I know if my efforts are working?

Analytics! Using Google Analytics, you can track performance of pages, and compare that to other pages to see if your hard work paid off.

What is a Search Engine?

What is a Search Crawler?

What is a SERP?

On-Page SEO Basics

Headings

Headings are simply the headline of the page, and should be tagged with "h1". All pages should have an H1 heading, and that heading should identify the main topic of the page. H1 headings should aim to be unique to the entire cmu.edu domain when possible.
H1s like, “News” or “Undergraduate Program” are vague, and likely used through many departments and colleges throughout the CMU domain.
Better H1s describe the page topic, even if it may initially sound a bit redundant. For example, “Psychology Undergraduate Program” is a better page heading than simply “Undergraduate Program” on a psychology page.
In rare instances, it’s OK to use more than one H1 on a page, BUT in most cases, we should aim to use one page H1 heading, and using H2 is preferred for all subheadings underneath the main page headline.

Titles

Title tags are the html tags that say what the title of a web page is and are one of the most important ways that pages are identified by both users and search engines. You can normally see the title of a webpage in a web browser’s title bar, as well as SERPs.
Title Tags will appear in search results in the following ways:

Because the title of a page is considered to be the single most important identifier of what the page is about, great care should be taken to ensure the title contains the top keyword(s) associated with the page. Ideal title tag length is <50-60 characters for top pages, but can vary.
Page Titles are often an extension of a page’s H1 heading, with a bit of CMU branding added automatically. Title tags will often include “- CMU - Carnegie Mellon University” to help identify the brand within SERPs. By default, your site name and Carnegie Mellon University are added after the page title in the browser's title bar.
In the CMS, when you give a page a title (not to be confused with headings as addressed above) you will see it reflected in three areas: the side navigation, the breadcrumb links and in the title bar at the top of the browser window or tab.
Title tags should be placed in the header html and are not visible to the user on the web page itself. Every page without exception should include a title.

Meta Descriptions

Meta descriptions provide brief summaries of a web page, and are meant to give search engines (and users) a tailored snippet of what a web page is about in SERPs. Similar to title tags, meta descriptions should be placed in the header html and are not visible to the user on the web page itself.
A good meta description:

• Accurately describes the content of a page
• Is more than 70 and less than 155 characters
• Is unique, and not duplicated from another page

Best practice is to include a meta description on every page.
**Note** just because a page has a meta description doesn’t mean Google or other search engines need to follow it on SERPs. Especially in the case of long, detailed searches, search results may often take a snippet right from a page body text instead of the meta description.

Alt Tags

Alt tags are used on images to describe the image. This is used heavily for both SEO search results and site accessibility for being able to quickly read and analyze web images. It is recommended to include alt tags in all non-decorative images sitewide. Writing a good alt tag includes describing the image in detail, summarizing the key points of an image, as if it was being described to someone who can't see it.

Content: Writing with SEO in Mind

When writing/publishing web content, first consider the purpose of a new page, and what the intent of the page is. 
If the page is intended to navigate someone into deeper content, ensure that every page is properly linked and explain where you’re taking users to. 
If the page is informational, do you want the content to appear to people outside of CMU for any reason? If so, consider the following to help optimize for search: 

Be Comprehensive for Content Length

A good SEO written article is as long as it needs to be to answer any and all questions. There’s no set rules to follow for length, but the best rule of thumb is to ensure that whatever is written is comprehensive.

A quality, comprehensive SEO article is the compendium of helpfulness. The best SEO pages answer all questions related to a query, so that a viewer feels no need to continue their search for more information.

If a topic can be discussed comprehensively in 500 words or less, great. But more often, more is more.

Longer pages > more pages. Data shows overwhelmingly that users would rather scroll through a longer page with more content than load multiple pages. Sort pages out using subtitles, and try to keep as much data as possible on one page, rather than spread across multiple pages.

Try to provide all similar topics and answer as many related questions on the same page, instead of breaking things out across multiple pages, especially if things are closely related. Average time on page and engagement data indicates that viewers prefer to scroll through a long article over navigating through smaller pages. Anchor links can also be added to provide waypoints to jump through to content quickly.

Can web page content be too long?

Short answer: we probably won't run into an issue when content is too long.

Long answer: As long as the content isn't going off script or combining content unnecessarily, there’s not many instances where page content is too long.

There's no maximum word count for page length. What we should be aware of is load time for pages, but this is more for visual elements like images and things like that. Text-based content loads super quick.

Additionally, we want to make sure we stay on topic. It wouldn't make a ton of sense to group a department’s research and admissions pages together normally, because those are different topics, and can have different (but sometimes overlapping) audiences.

Otherwise, being more comprehensive provides people with more of a sense of completeness, especially where a bulk of our content intent is, which is in providing users with information. Instead of feeling like you need to navigate from one page to another to find the information you're looking for, providing a complete experience without having to leave a page or wait for another page to load is going to provide better engagement, better SEO rankings, and better long-term web traffic.

As far as content length is concerned, when you wonder if something is long enough or too long, ask: "Is there anything else that people might be wondering about, if this was the only page they visit?"

Backlinko did a study that showed that top results are on average 1,447 words long or longer.

Reading web content isn't as linear as reading a book. In many instances, when we look at a web page, we scroll through sections quickly—maybe we read the first bit of content, and then we scroll through the subheadings until we find the section of information we're looking for. Maybe after that we scroll down, or we scroll back up, etc., but if someone is looking for information, they may only be looking for a specific thing. By frequently using subheadings to break up content, it allows users to navigate web pages in a way that’s more navigable to more people.

Another consideration is the quality and length of similar content. If we're talking about a subject that doesn't have much information anywhere on the internet around it, we might be able to get away with a bit less content on a single page. However, if other results are longer and more comprehensive than ours, we should aim to beat them in every single way, including how thorough we are. More is almost always more.

Let's use Shakespeare as an example.

When you Google "Macbeth" the top 2 results are Wikipedia, and then this link from the Folger Shakespeare Library, which is the play.

It's the whole Tragedy of Macbeth in one page.

A few more results down, and we have MIT doing the same thing.

Now, this isn't the entire First Folio of his many works, but it is a full play, not broken up by Acts or Scenes.

Why are longer pages better? An example
Consider Instagram. How do you interact with Instagram? Do you click into something and wait for it to load? Or do you just scroll infinitely?
Infinite scroll is preferred in most scenarios to multiple shorter pages, assuming the content is all connected.

Practice E-E-A-T

EEAT is one of the main factors for consideration in Google rankings, and it pertains mostly to content. EEAT stands for:

• Experience. How much first hand experience does CMU have on the matter written?
• Expertise. To what extent does the author (lesser so CMU/the org as a whole) know what they’re talking about?
• Authoritativeness. How much is CMU an authority on this particular subject?
• Trustworthiness. How legitimate is content from CMU?

The last two questions are so much of a given thanks to CMU’s global prowess, they can essentially be disregarded. With that in mind, when formulating new content, aim to focus on specific experience and expertise in an area to convey points and dominate search.

Search and observe.

Before writing anything, quickly do some research. It does not have to be complicated. 
Start by Googling your intended topic. There’s a strong likelihood that any topic being written has already been written about before, and in the words of Pablo Picasso and this author,  “Good Artists copy, great artists steal.”

Consider the following questions when searching for an intended topic:

• What are the first few results like?
• What kind of websites are in the top spots? Other colleges? Big companies? No-name websites?
• How long is the content there taking top spots now?
• Are their video guides?
• What questions are being answered from top results?
• Images for the search results?

Click into top results (at least the top three-four) to get an idea for page organization, topics you may want to consider, and the quality necessary to compete and beat what exists before publishing a page.

Keywords, Research and Keyword Inclusion

A very important factor to making your website search engine friendly is to populate your content with keywords that are highly relevant. Sift through your content and take note of important words and phrases that you feel are most relevant to the topic being discussed.

Overall, a website that has been optimized to use relevant keywords will have a greater chance at appearing in search results when those keywords are searched for. Keep in mind, your goal is not to just draw traffic to your website, but to draw targeted traffic (i.e., traffic that is specifically looking for the information you have to offer).

Before you start Keyword research:

• First, have a goal.
• Second, have an audience you’re creating content for.
• Third, have a topic you want to create content about.

You can’t perform Keyword research without having a goal, audience, and topic.

1 main KW + 2 related KWs +

For every new page, aim to have one main target KW, and two related keywords. Anything beyond that is a plus in most instances. Three is a good number to start with.

How many times should I mention a Keyword?

Write naturally, and don’t worry about keyword stuffing. Keep content simple, authentic, and helpful to a reader.

Keyword density is a term that refers to how often a KW is used in content. This has previously been a factor for search results, but in more recent algorithm updates, KW density has been largely replaced with more natural, authoritative writing.

Use Synonyms

Not everyone searching necessarily has the same lexicon of word usage as the author. For this reason, try to scatter synonyms throughout content.

Don’t Keyword stuff

Don’t put KWs into articles just to have them in there. The best content is expansive and KWs are used naturally.

Include KWs in headings

What about headings? See the headings section above, but aim to include Keywords in headings to outline the section below.

Create Keyword lists

Start with your intended keyword for an article, as well as synonyms. Then, ask yourself (and maybe search for) the following:

• What similar generic words refer to this focus area?
• What might someone search for at the beginning of their journey?
• What might someone search for after reading your content?

Use AI to create/check KW lists

Using ChatGPT/Bard/Copilot, enter something like the following prompt: ‘I want to write an article about {insert topic here}. List the top few KWs to include for best SEO results. Run the same prompt into multiple GPTs, aggregate the results, and search the KWs they include to double check accuracy.

Linking best practices and recommendations

Sticky Webs Catch the Most Flies. More links to more resources across the CMU domain, through the various forms of navigation, provides better experiences for both the user, and for search engines to help navigate a website. Relevant links in the body copy of text, in addition to top nav and side nav menus, is preferred for SEO and user experience. Contextual linking also helps avoid orphan pages. Aim to include at least 1-2 links per 500 words.

Use Linking Funnels

Links should be relevant to the content on the page. Simply put, find related content you think someone reading would want to read next. But how?

• Something higher up - If someone was reading this and felt like they were missing the first chapter, what link would you send them to?
• Something similar - If you like what you’re reading now, what else would you like to read?
• Something deeper - If you’re very interested in learning a lot more about the topic, what is the next level to provide someone?

Having trouble finding related articles? Google it! Read what CMU already has using ‘site:cmu.edu X search’

Avoid orphan pages

Orphan pages exist where there is only one link to a web page. This makes it difficult for both users and search engines to find and serve content, and should be avoided at all costs. Best practice is to either:
A - include more than one link to new pages
B - mold new content into an existing page, instead of creating a whole new page for a topic

**Note** Orphan pages aren’t orphaned if they are also included in HTML sitemaps as part of a cohesive web link structure

Use Absolute Links

Given the size of cmu.edu, best practice for links should be to use absolute links when linking across the site at all times.

Absolute vs Relative explainer: absolute links means using the full URL, relative means the web server is implied. For example:
Absolute link - <a href=”https://www.cmu.edu/news/”>
Relative link - <a href=”/news/”>

Between the various CMSs and subdomains, best practice is to use absolute links as a standard will help crawlers better understand the cmu domain.

Force https://

https:// is the standard. There are no exceptions, and we should never use http:// for internal links. This is imperative to force domain-wide. While redirects are in place to update this, it is still best practice to ensure all links use https:// when typed out.

Avoid Duplicating Content. Consider rel=“canonical”

A rel=”canonical” indication on a link informs search crawlers that content between two pages is either identical or similar, and should be attributed to the canonical link, instead of being considered duplicate content. CMU should consider the use of the rel=”canonical” linking when a page is heavily duplicated from an original source elsewhere within the cmu domain.

Example: Let’s say the news team was given a story from the SCS team about a recent professor’s achievement to include in the main news website. Instead of copying the content and publishing, the rel=”canonical” tag can be used on the new page to inform search crawlers that it is not the original news page, and the original is on the SCS site.