Mar 23, 2015
Keywords: You're Doing It Wrong
There’s a bit of confusion about how we should be using keywords today.
A lot of people believe that keywords should be included as many times as possible on the actual page; and in the meta tag named “keywords”, so that Google can read it, and find out what the website is all about.
While this actually *was* the case when search engines were first created, in the last decade or so, the sophistication of the algorithms means that now the meta keywords tag is not used at all.
There’s also an outdated notion that keyword density tells search engines that a page is exactly what a user was looking for. So a lot of people overuse exact keywords, which can result in content that’s barely readable.
If you’re using the keyword repeatedly, rather than pronouns, the copy may read something like this:
“Granny Smith apples are great in Granny Smith Apple pies. Granny Smith apples are a member of the rose family, and Granny Smith apples are bright green. Buy Granny Smith Apples at grannysmithapples.com.au.“
Another idea that sometimes comes up is adding all the keyword information for your site to the footer, and either leaving it visible, or hiding it with styling so that users can’t see it.
As it happens, concealing your overused or irrelevant keywords is actually a worse idea than making them obvious.
Keyword Stuffing Can Lead to Penalties
I’d like to think that we’ve moved away from the days of irrelevant keyword stuffing (here’s a great post by Matt Cutts, Google’s head of webspam, that talks about the unusual link between an immortality device and the tupac kazaa hospital), but on-topic keyword stuffing can still completely negate any positive effects you’re trying to get out of SEO.
Google can lay out two sorts of penalties if it decides you’ve done the wrong thing: manual penalties and algorithm penalties.
A manual penalty means that you will need to fix your errant content, and mea culpa in hand, send an explanation and description of your fixes off to Google who will review your site.
An algorithm penalty is potentially harder to come back from as you can’t resubmit for reconsideration (and you may not get a notification in Webmaster Tools). The penalty will not be lifted until the site is recrawled, reindexed, and processed.
The penalties themselves can mean that the whole site is dropped in the rankings; or that it’s dropped for a particular keyword, or a particular page. In any case, it’s a completely negative outcome for any work you’ve put into SEO.
The evolution of keywords
In the olden days, people would fill their on-page data with as many keywords as possible, and this actually did result in their sites getting a better ranking on the SERP (search engine results page).
However, businesses started trying to manipulate the system and began using either “adult-themed” or barely-readable combinations of keywords and copy in order to move up the rankings – and with roughly 30% of users clicking on the first result, you can see the appeal.
But the search engines began to realise that these sites might not be offering the best content – they were often spam sites themselves.
As a result, they improved their algorithms to ignore (or even penalise) the elements that can be easily ‘cheated’, and focussed more on elements that seem to measure user satisfaction.
In any case, trying to manipulate the search engines for particular keywords is a very tough job. Some very popular and competitive terms may be ranked for worldwide brands, and ranking higher than those would take years (if at all).
It helps to keep in mind that only 30% of searches are for these one and two word keywords. The other 70% is made up of long-tail keywords, where people are using more terms or queries around those phrases.
Google’s Hummingbird update
In August 2013, Google changed how they rank their sites with an algorithm called Hummingbird.
Hummingbird update main aim was for Google to really understand a user’s query (and there is some discussion that it’s due to more people using voice queries on their phone, so it’s more conversational).
It looks at each of the words someone uses in a query, and tries to match its actual meaning to particular pages.
One of the ways is does this is through something called TF-IDF (term frequency–inverse document frequency) – this tries to gauge importance by measuring how often a keyword appears in one document against how often it appears in many other, similar documents. If it shows up relatively often in one page, but doesn’t appear in other, similar, pages, that means the word is important on that page.
Another way Hummingbird works is by looking at “latent semantic indexing” (LSI). This again considers a collection of pages, and finds keywords that are often mentioned together. The pattern shows that those terms must be somehow related.
For example, Google would “learn” that the keywords flour, baking, and bread are semantically close, and a search for baker would bring up pages related to the keywords as well (use the ‘searches related to’ section down the bottom of the search engine page to find related LSI keywords).
What this all means is that the search engine is smarter about the information that a user is searching for.
The importance of click-through rate
But one of the most crucial uses of keywords is in the content itself. If a site ranks highly for a keyword or phrase, but doesn’t satisfy the user (who backs out and returns to searching), the search engine eventually learns that the particular site doesn’t seem to meet the keyword’s needs.
Along with Google’s PageRank, click-through rate and “long click” are two of the most important factors (of the ~200 in ranking measures) in how Google ranks a site. This great quote from Stephen Levy’s In The Plex explains it well:
"… Google could see how satisfied users were. … The best sign of their happiness was the "long click" – this occurred when someone went to a search result, ideally the top one, and did not return. That meant Google has successfully fulfilled the query. But unhappy users were unhappy in their own ways, most telling were the “short clicks” where a user followed a link and immediately returned to try again. "If people type something and then go and change their query, you could tell they aren’t happy," says Patel. "If they go to the next page of results, it’s a sign they’re not happy."
So, how can you use keywords, long-tail keywords, and latent semantic indexing to your advantage?
Where to use keywords: Content
The first step is to write your content.
Decide on your topic, and write on it well. Be engaging, or funny, or serious (tone is important), and check your grammar and spelling.
Then go back, read through, and make sure that your keywords are included. If they’re not, it’s likely that there’s confusion about either the keywords you’re targeting or your content isn’t relevant to your product or service.
Think about the questions that users ask related to your product or service. Ask your customer service team about the most frequently asked questions. Write content that answers these questions, and include your long-tail keywords and related LSI keywords where relevant.
Related reading: 11 Proven Ways to Find Content Marketing Ideas
And don’t forget to include the more general keywords that can set you apart from your competitors, both in a business sense, and a ranking sense.
In his post on Backlinko, Brian Dean talks about keyword classes around commercial intent. He suggests attaching words to your long-tail keywords like buy, coupon, discount, and deal.
For product reviews, try top 10, best, or review (obviously ensuring that the content is, in fact, a top 10 list, or a review). For blogs, or more informational content, think about the phrases that you would search for: how can I, I want to, best way to.
Landing pages (pages that can stand alone as an entry point to your website) are ideal mediums for this sort of interesting, relevant information.
Focus on your content, and build around a topic, rather than a keyword; this improves the experience for the reader, as well as for search engines.
Where to use keywords: on-page elements
Apart from content, the other on-page elements that are important for keywords collectively make up about 3.24% of Google’s ranking factors. But they can have a big impact on how a user approaches a site.
Keywords in a human-readable URL can attract users in search engines, email, social media, and anchor text.
A page’s title is what appears in the blue link text on the search results page, and again, can determine whether or not a user thinks a site can help. The meta-description is used in Google’s snippet, which is an important factor for the user in deciding whether they should visit a site.
Ensuring that you have filled in the URL, page title, meta description, H1, H2, and alt tags with content that’s descriptive, unique, and relevant will go a long way to optimising a page for both search engines and users.
If you’re looking for guidance on your SEO strategy, call Hayley on 1300 780 566 or send us an email.
How to find keywords
There are whole realms of the Internet devoted to keyword research, and I’d rather let those who’ve come before me repeat themselves, so please have a look at the following.
Brian Dean from Backlinko suggests *not* starting with Google Keyword Planner, as it can be very narrow.
Joey Barker from unfunnel likes Google Trends, but advises to only choose keywords that are trending upwards.
UberSuggest is popular for long-tail keyword research.
And Justin Deaville from Wordtracker suggests the tried and true Webmaster Tools for harnessing your own data.