Playful Search: 2. The Sea Analogy
In the previous article about How Search Relevance Works, I introduced the concepts of query intent and results diversity, and you learned how understanding the query is the first step to evaluating relevance. In this article, I will add two other elements to the analysis: the amount of content provided by the result and its quality.
The allegories, concepts and techniques presented on this Playful Search series are adaptations from the SEA Model, a course that I created as an alternative method of explaining search and result relevance to search quality raters. While the purpose of the SEA Model is to teach search evaluators how to come up with the most appropriate relevance rating for search results, the purpose of this series is to educate the reader about search and relevance so that you can use this knowledge to develop a stronger and more sustainable web presence.
For this article, we will begin with an exercise of imagination.
Think of the web as if it were the sea, where the searcher is the fisherman, the query is the bait and search results are fish. We are sailors, observing and learning as the fisheries unfold.
Based on our observation, we notice that every time the fisherman cast a different bait into the water, in other words, every time the searcher issues a new query, a specific shoal of fish — search results — is attracted.
In this figurative sea, big fish tend to show up first, for they are stronger, followed by medium and small ones.
Allegorical sea representing a web search process
Following that analogy, we can say that every search leads to a fish competition for the bait where big fish represent the potentially most relevant results — potentially stronger results — while the small ones represent the least relevant. Medium fishes have a potentially intermediate relevance rating.
Allegorical sea where different fish sizes represent different levels of search result potential relevance
In the previous article, you learned that when the query has many possible intents, it’s helpful to divide these intents into three categories — more likely, less likely and very unlikely intents. I mentioned that the most relevant results address intents that are more likely. But there are still some pieces missing in this story.
Introducing comprehensiveness
The idea that the most relevant results address the most likely intents of the query is correct, but incomplete. Besides addressing the correct or “more likely” intents, potentially most relevant results also provide an amount of content that is satisfactory1 for the query intent they are addressing.
(1) Satisfactory means that the amount of content provided meets or exceeds the user’s expectations for the query intent that the result is addressing.
Allegorical fish representing potentially most relevant search results
As you can see in the above image, the analogy of search results as fish is opportune to illustrate this relation between intent likelihood and amount of content, and how it affects relevance.
If the result addresses a more likely or correct intent but has an amount of content that is acceptable2 but not satisfactory, then it’s only potentially somewhat relevant.
Likewise, results that address less likely intents are also just somewhat relevant at best, no matter if the amount of content provided is satisfactory or acceptable2.
(2) Acceptable means that amount of content is reasonable, but not adequate, because it doesn’t meet the user’s expectations for the query intent the result is addressing.
Allegorical fish representing potentially somewhat relevant search results
Now, if the result addresses an intent that is very unlikely, then it’s potentially least relevant for the query, independently of the amount of content provided. The same is true for results that provide an inadequate3 amount of content. They are always least relevant, no matter the likelihood of the intents they are addressing.
(3) Inadequate means that the amount of content is not acceptable, because the user would be expecting much more from the result for the query intent that it’s addressing.
Allegorical fish representing potentially least relevant search results
Results that do not address any possible query intent are, of course, not relevant — they are fish that shouldn’t be attracted to the bait.
Being a big fish on the web-sea
If big fish are the potentially most relevant results in the web sea, and the concept of relevance only makes sense in the context of a search, then it’s impossible to be a big fish all the time!
A result that is very relevant for certain queries is just somewhat or not relevant at all for others.
Allegorical sea where the same web page has a different level of relevance for different queries
Therefore, the first step in being a big fish is to determine the best keywords to target.
From there, you need to make sure that, besides addressing the correct or any of the more likely intent/s of the keyword, the amount of content you are providing meets or exceeds the user’s expectations for that intent.
The Health Factor: Quality
If searching the web is like fishing in the high seas, in order to explain which fish will win the race for the bait, we will need a scale that is more granular than our three basic levels of relevance. This is where quality comes in.
Think of quality as the health factor of search results. It won’t turn a medium fish into a big one, but it can make it strong enough to compete. And as a health factor, the lack of quality can also make the fish weaker, less healthy, and cause it to lose positions to smaller fish.
Take this query as an example: <chocolate chip cookie>.
The most likely intents of this query are (1) to look for chocolate chip cookie recipes, and (2) to look for a list of chocolate chip cookie recipes.
Whenever it’s not clear whether the user would prefer a single item or a list of items — in this case, a single recipe or a list of recipes — you should consider both intents as equally likely.
Now, let’s suppose you’ve just launched a recipe blog and are looking to rank in the top positions for the query <chocolate chip cookie>. You expect to accomplish this by providing a satisfactory amount of content for the intent to look for chocolate chip cookie recipes, which is one of the most likely intents of this query.
Do you think users, on average, will find your recipe as relevant as the one of that famous chef’s blog that has been successfully sharing recipes online for years? Probably not, and neither will search engines.
Among every potentially relevant result, there will always be some fish that are stronger than others. And for most queries, it takes an adequate level of Quality to compete for the most prominent positions in each of the relevance ranges.
Above all, it takes an adequate level of Expertise, Authoritativeness and Trustworthiness (E-A-T), which is a major component of quality.
Remember: EAT is core to result quality, the health factor.
Makes sense, doesn’t it?
User expectations for amount of content
User expectations change over time, and also depend on the type of content provided by the result.
Here are some examples:
For a recipe page, displaying the ingredients and instructions might have been enough to meet the user expectations when looking for recipes in the past. However, recipe websites are now much more evolved, and most users expect that the page will also include at least some images of the dish, prep time and reviews from people who made it. And if you want to exceed their expectations, including an explanatory video, nutrition facts and additional tips might be the way to go.
For a product page, users with shopping intent are expecting at least an image, the product description, and other information that might be relevant for that specific category of product. A page that sells an electronic gadget, for example, would need to include the product’s features, specifications, and, ideally, user’s reviews.
If you are selling a type of product where design is very important, like clothing and footwear for example, then a user with shopping intent is expecting images that show different angles of that product.
When assessing the user expectation for amount of content, it’s important to understand that each intent brings its own specificities and requirements.
How to assess what the user is expecting
To assess what the user is expecting, put yourself on their shoes and, whenever appropriate, look at what top ranked results are delivering for the intents they are addressing.
It’s not my purpose here to digress on how Google might be assessing these things algorithmically, but it’s important for you to know that the prominent results of today can (sometimes) give you a hint of the amount of content the user is expecting for their intent.
If you are looking to create a video that targets the query <how to use chopsticks>, for example, it’s unlikely that you will be able to deliver a satisfactory amount of information in just 10 seconds. Most of the top ranked videos on this topic take at least one minute to convey their message.
Of course, there is always the opportunity to get creative and provide something different. But don’t forget that putting yourself in the user’s shoes is the most important thing to do when you’re looking for answers on how to satisfy their intent.
When targeting more likely intents, monitoring big fish that are currently ranking well is very important, as those results often end up shaping the user’s expectations over time.
To illustrate how this happens, let’s pick the recipes example again. Since most of today’s top recipe websites display images, prep time and reviews, then this becomes the standard — something the user is expecting when looking for these types of results.
The platform you choose impacts the amount of content you offer
Providing a satisfactory amount of content depends not only on the creator’s effort, but also on the platform’s features.
If users now take for granted that product pages should include reviews, then this feature needs to be available on your eCommerce platform of choice. Although this is not necessarily required to meet the user’s expectations for content when they are going to buy a product, such feature will contribute to increasing the amount of information you are providing for the shopping intent.
If you are looking to create a course that includes quizzes with logic jumps and calculations for example, make sure you opt for a platform where these functionalities are made possible, either natively or via integration. Switching between platforms will be difficult and, most likely, not be worth it once the course is already published.
Choosing the right solution is a crucial part of a successful content strategy. Be aware that this process may trigger some trade-off, and when deciding for a platform you will be opting for some attributes and features while sacrificing others, at least in the short term.
This is one of the reasons why it’s so important to choose the right businesses to partner with. In many circumstances, you won’t have all the features and resources you’re looking for at first, but if the provider is driven by a passion for excellence and innovation, then you can rest assured you will be in good hands for your solution needs in the foreseeable future.
Bigger fish may show up and you must adapt
If the amount of content you are providing is satisfactory for an intent today, that doesn’t mean this is going to be the case in the future. New content is being put on the web all the time and it’s inevitable that bigger fish will show up in the sea for many queries.
To succeed in this environment, adaptation is the key. Monitoring the shoal (i.e. the SERP) over time is a good habit to develop, and you must take the action to update your content when you find it necessary.
Updating the content can mean anything from including the latest developments on a given topic or new features on the page, to even excluding information that no longer contributes to meeting the user’s expectations for the targeted intent.
Remember: fish are living things, and content that does not evolve is fated to lose relevance with time for certain queries.
