Chapter 3. Writing Relevant Content for the Web Audience
Knowing what you know about how Web users find, scan, and ultimately use Web content, how do you create relevant content for your target audience on the Web? That is the question we will begin to answer in this chapter. The complete answer to this question is the subject of a whole book. And even then, there is no formula for every challenge you will face in connecting with your target audience. But we can lay a conceptual framework to help you make intelligent Web content decisions. We begin laying that framework in this chapter.
Providing relevance for your target audience fuels your writing. It determines what you write, how you write, and how you design Web content to cue into Web users’ needs. Relevance is the coin of the Web realm, and that is not in much dispute. But despite the general understanding of this point, the practice of creating relevant Web content leaves much to be desired. Why? The main reason is that the medium is relatively new and continually evolving. Many site owners assume that there is not much difference between writing for print and writing for the Web. So they apply the same processes for the Web that they do for print publications. But the practice of writing for the Web is so different from print writing that their content is often not relevant to the users who find it, if users find it at all.
Suppose you assume, like so many content owners, that writing for the Web is similar to print in many ways. You develop a topic structure and flow not unlike the chapters of a book. You consider how users might follow your flow sequentially from topic to topic. And based on this topic flow, you assume that users know topics at higher levels of the topic structure if they are on topics at lower levels. In short, your content planning resembles planning book chapters, except that you might allow users to skip topic pages they are already familiar with. According to this scheme, if users find your home page relevant, they should also find the related pages relevant.
This is not a bad practice in and of itself, but it ignores the most important aspect of Web publishing. Namely, it does not take into account users who come to your pages from other sites, especially search engines. If other site owners find your content useful (your sincerest hope), they will link to it. And they won’t necessarily link to the top page of a topic hierarchy, or to your site map (the Web equivalent of a table of contents). More often than not, they will link to lower-level pages in your site—those that are most directly relevant to the points they are trying to make. In the case of search engines, they will link to the pages that are the most relevant to the keywords that users enter into search fields. The results will be based more on the semantics of your pages than on where your pages are located within your site architecture.
We suggest that you should assume that most of your users come to your pages from search engines. If they navigate to one of your pages from others in your environment, you know that the links they click will lead them to relevant experiences, because you control those information paths. Consider that challenge solved. But if you can develop relevant content for users who come to your pages from external sites, especially search engines, you can dramatically increase the overall relevance of your content for your visitors. Search use has increased fairly steadily since Google launched in 1999. It will never entirely replace navigation, and other kinds of external referrals will always be present. But developing relevant content for the search audience will do more to improve the overall relevance of your content than any other Web writing or site architecture practice.
Who Is Your Web Audience?
In print contexts, the writer controls how content is conveyed. If readers want to understand print writing, they must follow the flow of information that the writer presents. You might not want to read the words in this chapter, but if you want to understand the rest of the book, you need to do so. As book writers, we have the luxury of presenting information in the way we think is best. Though we want to be sensitive to the audience, we honestly don’t know them, and therefore, we mostly write the way we think is best and hope they understand us.
In particular, some of the information we present here might be irrelevant to our audience (you), either because you already know it or you just want to skip past the conceptual stuff and get to the how-to part of this book. We know you’re interested in writing, so you will be inclined to give us the benefit of the doubt if the information is relevant to that. But we don’t know your time constraints, patience level, or existing knowledge. So we present the information that we think will serve most of our readers in the best way we know how, and hope you will cut us some slack if we don’t address your needs perfectly.
If this were a Web text, we could present the information in a more modular way, giving you all kinds of contextual cues about what you might skip, and about what everyone who reads this text needs. We could also use knowledge about how users find parts of sites though search, so that we could attract the people who are interested in the information in some of the modules. And we could look and see how users are clicking on a site to determine what is more popular and what tends to be skipped. On the Web, we can know all these things about how content becomes relevant to our users. In print, one thing we can do is to supply graphics and sidebars to help you skip the parts of this work that you don’t need. But the static medium is a poor substitute for properly presented Web information.
When you write content explicitly for search engines (so that your users will find that content), in a sense, you invoke them with a compelling mix of keywords and links that draws users to your pages. The challenge is to craft your pages in ways that attract specific users from search engines, especially Google. In so doing, you can present relevant information to your audience. This book is about attracting your audience with keywords and links and thereby providing relevant information to them. As mentioned earlier, because search engine results cater to the way users scan and retrieve content on the Web, writing for search engines is also an effective way to write relevant content for your audience.
What Is Relevance on the Web?
The study of relevance is messy and complicated. But there’s no running from it. If you want to be successful in writing for Web scanners and readers, you have to come to an understanding of how these people determine relevance on the Web. Like other aspects of the distinction between the oral, print, and Web contexts, how relevance is determined depends on the medium. But there are some common elements to the way relevance is determined in general. Though this book is about relevance on the Web, it’s instructive to consider the more familiar oral and print contexts. We can then use comparison and contrast to show the different ways relevance works, depending on the medium.
The relevance of a statement, sentence, or fragment is determined by the context of the communication that surrounds it. In oral contexts, this means that what was said prior to a statement in part determines the relevance of that statement. In print contexts, the relevance of a sentence is determined by the sentences and paragraphs surrounding it. And in Web contexts, the relevance of a piece of content is determined by the elements of the page on which it appears, and by how that page relates to other pages, both internal and external. For a user, the context of a Web page is the last few pages he or she visited before visiting it. The page that referred a user to your page will give the user the strongest sense of your page’s relevance. Something two clicks before your page will give a lesser sense, and so on. But because the user could come to your page from just about any other page on the Web, only some of which you have knowledge or control over, relevance is a particular challenge in Web publishing.
So what happens when a user comes to your carefully planned site from another site, such as a search engine? All the assumptions you have made about what is relevant to your users will not hold true. That user will not have followed your hierarchy of topics from the highest level to the lowest level. The user will land on your page cold. Unless your page can stand on its own for this kind of user, he or she will not find it relevant and will bounce off it. This often happens because many pages cannot stand on their own—they depend so crucially on other pages in their Web site’s hierarchy that they make little sense to users who come to them from search engines or other external links.
Like it or not, most users use search to find what they’re looking for on the Web. It is simply the quickest and easiest way to zero in on the most relevant information. Try as you may to control their activities and give them necessary background information through clear site architecture, they will ignore it and go straight to the answers to their questions or the solutions to their problems. Even those who come to a page on your site from its home page will often use your site’s search function if the site is at all complicated. If your site search doesn’t give them relevant results, they will often revert to Google. When Web content producers fail to produce relevant content for their users, it’s generally because they didn’t realize the vital importance of search on the Web.
As we indicated in Chapter 2, the crucial role of search is the root of the differences between print and Web media. Each page on a site must be able to show search users that it is relevant to them. This leads to a simple yet powerful tactic, which can turn ineffective Web sites into effective ones: Develop your content to be relevant for users who land on your pages from search engines, and it will be relevant to the majority of your users. The converse is also true: If you do not take search users into account when you develop your pages, the majority of your users will not find your content relevant.
Defining Relevance in Terms of Search
A good question to ask here is: What is relevance? For a piece of content on the Web, relevance can be defined as it applies to search engines. In this book, we focus on Google because it is the most used search engine on the planet. Bing is a good search engine that is gaining some traction in the search market. Baidu is the leader in China. Other countries have local search engines. But Google is the clear market leader, so it has earned our focus.
Perhaps Google’s own definition of relevance can help. Google has put its quasi-definition of relevance on its site:
That which has a logical or rational relation with something else, such as with the matter under discussion or a claim being investigated, etc., or has some bearing on or importance concerning real-world issues or current events.1
1 Taken from www.google.com/notebook/public/07491261619920101511/BDRreIgoQ853VyIsi. We call this a “quasi-definition” because it has numerous problems of definitional form. First, this definition is circular. Second, it does not encapsulate the meaning of the word relevance, but only gives examples of relevant things. Third, some of the examples treat relevance not as a relational concept but as an absolute (“concerning real-world issues or current events”), which violates its own definitional criterion.
This is not very helpful. It says what we have already indicated: Relevance is a relational concept between a linguistic item and its context. For example, two sentences are relevant to each other if they’re logically related, just as scientific evidence is logically related to a conclusion. But this is still not helpful. Relevance for writers is not about the semantic qualities of two pieces of text. It is about crafting content that will meet their audience’s needs. By removing the individuals from its description, Google removes the central aspect of relevance for writers—the reader/writer relationship. In short, though the relational concept discussed in this definition gives us hints, it doesn’t give us much understanding.
A better approach is to describe cases involving search and see what makes a page relevant to a user. To do this, we need to give an example of a success and an example of failure. We define success as having a user land on a page that he or she deems relevant and then become engaged with that page—by clicking something. We define failure in terms of bounce: If a user bounces off a page within a few seconds, he or she finds the content irrelevant, and the page has failed.
Given the above definitions, here are examples of success and failure.
Success: Suppose a user enters a phrase—say SOA—into Google’s search field and clicks Search. When the user gets the search results, she clicks one on the first page and lands on a page that she deems relevant to the search query—say, IBM’s main page on Service Oriented Architecture. The user then indicates relevance by clicking one of the calls to action on the page.
Failure: A user likewise engages Google’s search application by entering SOA into the search field and gets pages of results. Clicking a result on the first page lands the user on the same IBM page, but this time the user finds the page completely irrelevant. He was looking for pages related to the Society of Actuaries, not Service Oriented Architecture. Having failed to read the short description of SOA from IBM on the search results page, he clicked the result anyway, landing on the irrelevant page. The user then indicates that the page is irrelevant by immediately clicking the Back button.
Before refining our definition, we don’t want to lose a key insight from this example. Because there are so many different meanings for the same strings of letters (e.g., SOA), Google often can’t divine what a user means by a particular search string and must make what amounts to a best guess. It does this by seeing what use its users tend to make of the words they enter into search strings. Google tracks whether users bounce off a page or engage with it. If its users bounce from pages that define SOA as Society of Actuaries more often than as Service Oriented Architecture, Google will guess that the majority of users mean Service Oriented Architecture when they type SOA. This is not only true of acronyms and other abbreviations, but of whole-word synonyms. Google makes similar probabilistic decisions based on a host of linguistic complexities in natural language. For this reason, its algorithm is not just the semantic concept that we quote above. It also includes all kinds of contextual effects, such as links, which determine how its users use the Web to communicate. We will return to these complexities later in the book.
Note that requiring engagement narrows our definition of relevance. A user might land on a relevant page and bounce off because she had already consumed all the information on it. So, although the page was relevant to her, it was no longer relevant enough to be worth her time. So she then clicks the Back button. This example demonstrates an important part of relevance that no purely semantic concept can capture: Relevance is not an absolute concept. Items are more or less relevant to users. Whether they choose to indicate with their clicks that a page is relevant to them is a matter of whether it is relevant enough. Web users are typically impatient, and their patience is variable. A page must be strongly relevant to a user to entice her to click one of its calls to action. That’s why we narrow the definition of relevance considerably when we require engagement in our functional definition.
We will put the gory details of the linguistic underpinnings of relevance in a sidebar. Those skeptics among our readers who are inclined to dismiss our views because we use a functional definition can consult the sidebar (and the texts it refers to, if necessary). Those willing to give us the benefit of the doubt only need to know that the study of relevance is part of the field of pragmatics, the part of linguistics that deals with contextual effects, with how the semantic value of elements such as sentences and statements change from one context to another. Dan Sperber and Deirdre Wilson wrote the definitive text on the subject of relevance—aptly called Relevance (1986). It forms the basis for the following definition of relevance (adjusting their terminology to suit our needs).
- A piece of content is relevant in a context to the extent that its contextual effects in the context are large.
- A piece of content is relevant in a context to the extent that the effort required to process it in the context is small (Sperber and Wilson 1986, 125).
The only thing left is to explain in layman’s terms what Sperber and Wilson mean here. First, consider that this definition contains extent conditions, which state the realm in which something is relevant. In other words, relevance is not an absolute either/or consideration, but a matter of degree. Second, the first extent condition is positive; it defines to what extent a piece of content is relevant. A piece of content is relevant to the extent that it causes a large change in the audience. The second condition is negative; it defines how relevance is limited by matters such as time, space, attention, and patience. A piece of content is relevant to the extent that it requires little effort (time or attention, for example) to grasp it and act on it.
But what is a contextual effect? It is a change in the audience. Sperber and Wilson define contextual effects in cognitive terms, that is, in terms of the audience’s mental state. If the audience has an epiphany, that is a large contextual effect. If the audience merely tweaks one of its beliefs, that is a small contextual effect. However, for our purposes, we will not talk about the audience’s cognitive states. Instead, we will only talk about audience behavior, because that is the only thing we can measure. It is likely that every click from a user results in some change in her brain state (and some users click mindlessly). But we are nowhere near smart enough to know what that change is or how to measure it, especially in such complex creatures as humans. So, for our purposes, concentrating on functional/behavioral contextual effects will do.
So far we have defined relevance in terms of user interactions with Google. You might think this odd. We could easily define it in terms of how different pages relate to each other in a site’s architecture. There are any number of other ways we could have tried to define it as well. Why use search as the central concept? One reason we have already given is that many users start looking for information at Google and other search engines. So this seems a likely place to start. But there is another key reason that Sperber and Wilson’s work suggests: All things considered, the less effort it takes to find the information you need, the more relevant it is. These two reasons are related: Google is popular because it tends to produce relevant results in Sperber and Wilson’s sense. That is, more than any other Web application, it tends to help users find content with the least effort and the largest contextual effect.
So our choice of search in general, and Google in specific, is not arbitrary. Google is the bellwether of relevance: If you want to create the most relevant content for Web audiences, write for Google first, and take care of all other aspects of effective Web content later. That is the central thesis of this book. But what does it mean? How do you write for Google first? Read on to find out.
On Writing Relevant Content for Web Users
We will go into the specifics of our Google-first methodology in later chapters. But let us sketch the approach here.
In most publishing environments, we decide what and how we want to communicate with our target audience. We design messages that we think will engage that audience and inspire them to act in ways we desire. We then define successful communication by what percentage of our target audience acts the way we hope it will (such as purchasing ads, writing a paper that complements ours, purchasing a product, or writing a review). The Google-first methodology does not reject this basic process; it turns it on its head.
A common mistake in writing for the Web is to optimize content after the fact. Again, based on the model of crafting a message that you want to communicate and following the other steps in the chain, the last step is to prepare the content for the Web. Many content producers only consider search optimization after they have determined what they want to communicate. They take existing pages, do some keyword research, and map the pages to the relevant keywords that are most often typed in Google’s search field. Then they add those words at the appropriate places on the page.
The problem with this approach is it does not take into account that the content on the page doesn’t necessarily match the keywords very well. In some cases, the words you choose as you write happen to be the ones that searchers use. But that is rare. In many cases, especially in large corporate organizations with their own cultures and terminologies, the words that writers choose to communicate their messages do not match those that searchers use. If you optimize a page after the fact, the words you put in the title tag, the heading, and so on might not match the body copy of the page you wrote. Users may be attracted to one of your pages, but many of those who come to it from search will bounce off it when they determine that the content of the page is not relevant to their chosen keyword.
In extreme cases, Google will look at this after-the-fact process and see manipulation. Google has a filter for such efforts, and it will expunge your page from its index if you try too hard to optimize for content after the fact. One of the reasons Google is the search leader is that it aggressively filters out those who maniplate its algorithm by trying too hard to optimize pages after they are written (intentionally or not). This makes sense, because Google’s audience consists of search users. Google measures success based on the relevance of the results to the search terms that users type into its search field. If its users find the results of their search efforts irrelevant, they might try another search engine. Expunging irrelevant results from its index is a key tactic for Google to stay ahead of the competition. The other primary way that it competes is by determining that the chosen keywords for a page are tightly relevant to what the page is about. If you optimize after the fact, Google might still index your page and even display it to users, but it will not rank among the top pages for that keyword. Indeed, it probably won’t be anywhere in the top five pages of search results. For all intents and purposes, then, your content will be irrelevant to your target audience.
Unlike with other media, we do not have control over who consumes our messages on the Web. Targeting their audience precisely, PR folks send press releases to media representatives, marketing folks send direct mailers to an approved list of customers and prospective customers, sales people send proposals to known leads, and academics write for a known audience of journal subscribers. But on the Web, anybody or nobody might find your content in the first place. And those who do find it might consider it irrelevant on first glance. So attracting the target audience comes first. And, as we have demonstrated, the best way to attract the target audience is by optimizing your content for Google. This means performing keyword research before you even start writing, and determining what to write, how to write it, and how to deliver it to Web users based on your keyword research.
If you do this properly, you will not simply attract whoever comes to your site from Google and hope for the best, as many search engine optimists recommend. Instead, by proceeding correctly, you will attract an audience that is likely to engage in your content and perform the actions you desire them to. Part of the transformation from the print model to the Web model is adjusting expectations. Direct mail is considered a success if 1 percent of the audience takes the desired action. Magazine advertisers are often happy to get 50 new leads out of 50,000 readers. But on the Web, you can expect the majority of your users to engage with your site if you do your keyword research and craft a structure of pages to serve their content needs.
As you begin to consider turning the publishing process on its head in writing for the Web, consider the main benefit of starting with keyword research and developing pages only after you learn how the target audience uses language: Keyword research is a powerful way to analyze your audience. In print contexts, you might form a fictional picture of your target audience based on their backgrounds and assumed shared knowledge. But still, you are choosing to communicate with them using just your words. But you can never be sure that your word choices will match audiences’ word usage.
On the Web, you learn how your target audience uses language and you address them with their own words. Done right, keyword research can let you precisely bridge the gap between your messages and your audiences’ needs.
Summary
• The Web opens a new avenue for audience analysis. Unlike print, where we must form fictional pictures of our audience and guess how best to invoke this audience, by doing keyword research before creating our content, we can address our target audience with their very own words.
• For all intents and purposes, relevance on the Web can be defined by the behavior of users coming from Google. To the extent that users come to your site from Google and engage with your content by clicking its links, you can determine how relevant your content is to your target audience.
• These two principles lead to the central thesis of this book: The most effective way to create Web content is to take a search-first approach.
1. Define your target audience and perform keyword research to see what terms they use to find relevant content on the Web.
2. Develop a list of core words that your target audience uses and write one page for each of those words.
3. Test your pages to ensure relevance, and adjust as necessary.
• If you adopt a search-first approach for the Web, you can realistically increase your expectations about audience engagement. Whereas you might be satisfied with 1 percent engagement with some print efforts, you can shoot for 50 percent engagement rates or better on the Web. Unlike print, you can tune your pages to improve engagement rates as time goes on.