Based on work practice, the author introduces the relevant knowledge of information architecture for us to learn and reference together.
1. What is Information Architecture
Information architecture (information architecture), referred to as IA. It was born from the field of database design. When searching on Baidu, it gave a definition: the main object of IA is information, and the information architect designs the structure, decides the organization method and categorizes it, so that users can interact with the information. An art and science that is easy for users to find and manage. To put it more simply, it is a form of presentation of a reasonable organization of information.
So based on the above concepts, I think that the information architecture often said in the software industry is to design a content structure that is convenient for users to understand and find.
If we want to know what kind of information architecture is easy for users to understand and find, then we must first understand how users generally understand and find information content.
1. How and how to find
There are generally the following scenarios for searching, knowing what you are looking for, changing from looking for A to looking for B, wanting to get a certain type of content, and looking for things you have checked before again.
Know what you’re looking for. For example, Xiao Ming wants to go to the library to borrow a book today, but he doesn’t know the route to the library, so he opens the map software, searches the library, and follows the route to the library to borrow the book smoothly. This is typical of knowing what you are looking for in everyday life.
From finding A to finding B. For example, on the way home from the library, Xiao Ming felt that walking was very difficult, and suddenly found that he was too fat now and needed to lose weight, then he did not know how to lose weight more effectively, so he searched the Internet for related problems and found that running And swimming is more effective for weight loss.
He searched for relevant information about running and swimming again, and found that there was only one swimming place near his home, and the fees were high, and the park downstairs was just suitable for running, so he decided to give up swimming and choose running; so he went to search again about running. Precautions, suitable time for running, how to warm up, etc.
So, this is a typical change from looking for A to looking for B. At the beginning, Xiao Ming wanted to solve a certain problem, but he didn’t know how to solve it, so he kept iterating on the search problem by adding the information obtained and thinking. Originally, I just wanted to inquire about how to lose weight, but finally it became an inquiry about the precautions for running.
Want to get a certain type of content. Let’s continue with the example of Xiaoming. Xiaoming is about to graduate and is preparing for his graduation thesis. Because he has just been running recently, the weight loss effect is obvious, so he wants to write a thesis on weight loss.
Then Xiaoming searched for information about weight loss on the search website, and saved or cut all the useful content, so he learned that the current proportion of obese people, their troubles, most of them The weight loss method chosen, the weight loss effect, etc.; so Xiao Ming quickly completed the thesis through this information.
This is the scenario where you want to obtain a certain type of information, usually, multiple searches are made, using similar and related keywords.
Look again for what you’ve looked up before. Come on, continue Xiaoming’s performance, no, it’s an example: After Xiaoming finished writing his thesis, he found that there was a content about the current status of dieters that was missed during inspection, so Xiaoming also entered the search website and found it in the search records. This time, I searched for the keywords of this information, re-queried, found the same content as last time, and completed the information in this regard.
This is the scenario of finding again what was looked up before. Usually, the same keywords are used to query the same content and obtain the same information.
In the search scenario, the normal methods that people use are browsing, searching, and asking. This category is easier to understand, so I won’t introduce it too much here.
2. The way of understanding
Understanding, especially for the five thousand years of Chinese culture, is profound and profound. Different words, in different scenes, using different contexts, express completely different meanings.
Then the same is true of words. In different scenes, different words also express different meanings. So how do we understand it?
First, create a scene that has a sense of place. For example, when Xiaoming was writing a paper on weight loss, if he mentioned tea, the reader must think of information related to weight loss tea. This is the sense of place created in the textual scene.
Second, be modular. For example, if we are going to read Xiao Ming’s paper, then according to our conventional thinking, the title should be at the top. Below is the author information, the table of contents, and then the text. Therefore, in this process, we should pay attention to the location, structure, appearance, layout planning, etc. of each information.
In addition, we can also learn from the design of the physical environment. When I wrote this article, I happened to be in the library, I found a book, and then I learned that the first floor is to distinguish adults, teenagers, children, exhibitions, etc. The second is to classify books by alphabetical order. For example, IT books can be found on the bookshelves of Category I; and then the books on the current bookshelf are classified and explained on both sides of each bookshelf. For example, this bookshelf contains SQL, python and other books; then find the corresponding book according to the unique number corresponding to each book in the position corresponding to the bookshelf.
This can be understood as the design of information architecture, we can find them in many scenarios in our daily life.
2. Introduction to Information Architecture
Information architecture is generally divided into top-down architecture and bottom-up architecture.
The components of information architecture generally include: organization system, labeling system, navigation system, and search system.
It can help us browse the content of the website more clearly and conveniently with a preset path. We call it a browsing helper, which mainly includes: organization system, site-wide navigation system, local navigation system, site map/directory, index, guide, Site wizard, contextual navigation system.
It allows users to query string characters and displays a set of query results that meet the conditions. We call it a search helper, which mainly includes: search interface, query language, query builder, retrieval algorithm, search area, and search results.
The components that users need to reach their goals and view content mainly include: titles, embedded links, embedded metadata, blocks, lists, sequential helpers, and identifiers.
In addition to this there are invisible components such as: controlled vocabularies and thesaurus, retrieval algorithms, best match. The above content will be introduced to varying degrees below. I won’t go into too much detail in this chapter.
3. Organizational System
We organize information so that people can find answers to questions, provide scenarios for understanding those answers, and support casual browsing and directed search. This is the organizational system, which is composed of organizational plans and organizational structures.
1. Organizational plan
Let’s talk about the organizational plan first. We can understand that the plan for classifying and sorting information is the organizational plan. The organizational scheme is further divided into precise organizational scheme and vague organizational scheme.
The precise organizational scheme is very easy to understand. Commonly used schemes are alphabetical, chronological and geographic order, such as our mobile phone address book, which is sorted alphabetically; news in news software is sorted by time; information in map software is sorted by distance.
What about the vague organizational scheme? Commonly used are thematic organization plans, task-oriented organization plans, specific audience plans, metaphor-driven plans, hybrid plans, etc.
For example, the classification of various categories in Taobao is the theme organization plan; the classification of the top functional area in word is a task-oriented organization plan; the children’s mode in mobile phones is a plan for a specific audience.
2. Organizational Structure
There are two main types of organizational structures, one is a hierarchical structure, which is a top-down approach. There is also a database schema, which is a bottom-up approach.
The hierarchical structure is mainly about the mutual disassembly of information, but it is necessary to pay attention to its depth and breadth when disassembling each other.
If you do not consider the depth and there are too many levels of dismantling, then users need to keep clicking to view the final content when browsing, which is a very bad experience, so it is necessary to carry out wide and shallow dismantling. In this way, the information is placed on the main page, and the user can browse it and reach the final destination page with only one or two clicks.
The database schema mainly organizes information through the metadata in the database. In this process, the main consideration is the mapping relationship of metadata, as well as the relationship of content in hypertext.
4. Labeling system
The common label in daily life is to put a corresponding label on the edge of a certain page of the diary to indicate that this is the content of XX; it is common in software that there will be a label in some articles. the content of the label.
The topic of some software is also a kind of label. When publishing content, a topic is added; when we query the topic, we can see all the content posted by adding the topic. The above are all tags, but there are more than that. For example, the menus of websites or mobile software, or the icons at the top of mobile phones are all tags. The following is a brief introduction to the types of common tags.
1. Types of labels
Common types of tags are: tags for contextual links, tags for titles, tags within navigation systems, tags as index words, and tags for icons.
For example, when we are reading an article, the article may talk about a certain professional term, and this term can be clicked to jump to the page of the detailed introduction of this term, then this is the label of the contextual link. It is usually a hyperlink that appears in textual content to explain something or lead to some related content.
In our mobile application market, we often see certain categories, such as download list, free list, etc. These are tags that belong to headings, and are usually built with uniform numbering, font size, color, style, whitespace, and indentation, or a combination of these.
We open a website and see several menus at the top of the home page, including home page, product introduction, success stories, and about us; or open the app and see menus such as home page, community, and personal center; when clicking on a menu, several menus will appear. menu options. These are all labels within the navigation system, and its main point is to emphasize consistency.
For example, when I am on the homepage, the homepage is called the homepage, and when I click on the personal center, the homepage is called the homepage, which will cause misunderstandings to users and a bad experience.
Searching for a topic we mentioned just now, you can see all the content posted by adding the topic. This is the label as an index term. It enables precise search of content and is a valuable method.
When our mobile phone is muted, a small bell icon will appear on the top of our mobile phone; when we set an alarm, there will also be an alarm clock icon on the top of the mobile phone; this is a typical icon label. It is actually very common in life, such as the male and female logos at the door of the bathroom. However, in the use of icons, attention should be paid to the accuracy of the expression, that is, there will be no objection to the icon, and after the user sees it, the first reaction can accurately understand the meaning it wants to express.
2. General principles
When designing labels, there are some general principles that can provide us with good help when designing labels.
Narrow the scope as much as possible. When designing the label, the smaller the scope, the more clearly and effectively it can express its meaning.
consistency. As mentioned earlier, consistency is important in this process, and consistency includes not only naming, but also style, layout, grammar, granularity, comprehensiveness, and users.
In addition to the above information, one more important thing is to optimize and adjust your tagging system from time to time.
5. Navigation system
The navigation system is mainly divided into embedded navigation, auxiliary navigation and high-level navigation methods. There are 3 main types of embedded navigation that are included in most websites. Assisted Navigation It can provide assistive methods for finding content and completing tasks.
1. Embedded Navigation
There are three common types of embedded navigation, namely global navigation, local navigation, and contextual navigation.
Global navigation is a global navigation system that will be displayed on every page of the website; there are usually one or more local navigations, which we can call a sub-site or a website within a website, such as clicking a menu in the global navigation, The corresponding sub-menu will pop up or display, and this sub-menu is the local navigation; the contextual navigation can be a hyperlink in the article, or a title, etc. Click to jump to a new web page, download a file or specify an object.
2. Assisted Navigation
Assisted navigation includes sitemap, index, guide, configurator, search.
Sitemaps are more suitable for large-scale systems organized in a hierarchical structure. If the structure itself is not very hierarchical, it is recommended to use indexes or other visual representations.
Indexing For small sites, you can use your knowledge of the content to decide which links to introduce, and then create the index manually. For large websites, a controlled vocabulary can be indexed at the file level to automatically generate a website index.
Guides include guides, tutorials, and walkthroughs for specific users, topics, or tasks; configurators can be understood as a special form of guides; search is a core part of assisted navigation, which we’ll cover in detail in the next chapter.
3. Advanced Navigation
Advanced navigation is mainly divided into personalization and customization, visualization, and social navigation.
Personalization and customization can be customized for individual behavior, but the disadvantage is that it is not suitable for broader research; visualization is picture or icon type, such as a list of pictures when purchasing products; social navigation is mainly algorithm-based navigation .
6. Search system
The search system mainly includes determining the search area, selecting the components to be indexed, searching algorithms, displaying the results, and designing the search interface.
1. Determine the search area
First, let’s look at determining the search area. The search area is actually a subset of the information environment, and irrelevant content will be eliminated when searching. Commonly used methods are
navigation/target, indexing for specific users, indexing by topic, indexing recent content.
In the navigation/target, there are at least two kinds of web pages or screens, the navigation web page and the target web page. The goal of a navigational page is primarily to get you to a landing page, which holds the actual information you want to search for. In fact, for this kind of search, it is essentially a precise search.
Index a specific user. In layman’s terms, it is actually to understand several user groups faced by the search engine, and set targeted functions or filter conditions for the corresponding user groups.
Index by subject. For example, when I search for apple on an e-commerce platform, there are a lot of results, so indexing by topic is to filter the results again. I can choose a mobile phone, mobile phone case, computer, etc. This is searching by topic.
Index recent content. It is more common in life to search for news. Because news is time-sensitive, after searching, you can filter by time. For example, today, yesterday, last 7 days, last 1 month, etc. Or when searching for other content, you can choose to search for content within a certain period of time.
2. Select the content components to index
In this process, the main thing to do is to determine what content is searchable. For example, on Dianping.com, the searchable content includes business name, location, and dish name; only after the components to be searched are determined, there will be no confusion in the subsequent use.
3. Search Algorithm
In fact, search engines have many algorithms, and some have existed for decades. Because the search algorithm belongs to the core of the search engine, the pattern matching algorithm is mainly introduced here.
Pattern matching algorithm is one of the algorithms used by most search engines, and the main concern in this algorithm is recall and precision.
The recall rate can be understood as, when I search for a guitar, all the information about the guitar will be queried.
Then the precision rate will only query the content with high accuracy. The two are inversely proportional. That is, if you want to increase recall, you will decrease precision accordingly, and vice versa.
Recall = number of relevant documents retrieved / total number of relevant documents in the system
Precision = Number of Relevant Documents Retrieved / Total Number of Documents Retrieved
Other methods include, for example, document similarity, collaborative filtering, citation search, etc. I won’t do too much introduction here.
In addition to this, the query builder will also be involved. It is biased towards the backend and rarely faces the user. Its main role is to improve query performance. It contains spelling lookup tools, phonetic tools, stemming tools, natural language processing tools, controlled vocabulary and thesaurus.
4. Display the results
Regarding displaying results, here we discuss what components to display, how many documents to display, how to sort the listed results, whether to group the results, and how to act on the search results.
First, we need to consider which components need to be displayed in the search results. The general advice is to show less information to users who know what they are looking for and more information to users who don’t know what they are looking for.
Another way to do this is to provide representative content components, such as title, author, to users who know what they are looking for. Users who aren’t sure what they’re looking for can benefit from descriptive content components, such as summaries or keywords. If, for example, the titles are similar, more information can be displayed for users to distinguish.
Then we’re talking about how many documents to display. To put it simply, it is the search result page, and how many pieces of information are displayed on each page. This is mainly related to two aspects, one is the amount of information displayed by each result, and the other is related to the screen size we use. When we design it, we need to consider these two aspects.
Then talk about how the listed results are sorted. Regarding the way the results are arranged, it should be differentiated into sorting and ranking. Sorting is our common sorting by time or alphabetically; ranking is ranking according to relevance, ranking according to popularity, ranking according to user or expert evaluation, ranking according to paid position. Often they are used together, for example results are sorted by time and rating.
So after understanding the sorting, we need to consider whether to group the results?
When considering grouping, we have to consider the amount of platform content. According to common sense, the content of our platform will only increase, not decrease. When the content increases, there will be confusion in the content, and at this time, grouping is required. For example, in Douban, the search results will be divided into users, books, movies, groups, etc.
After we’ve grouped the results, there’s one more thing we need to think about, which is what we want to do with the search results.
In general, you can operate directly on the results. For example, in the app store, you can directly install or open the searched APP; in another case, you can select a subset of the results. For example, in an e-commerce website, the search results can be Add any item in the shopping cart, and then perform bulk operations in the shopping cart.
In addition, there is another easily overlooked, is to save the search content, not the results. On Amazon, for example, after a search is complete, you’ll see an option to save your search in the upper-right corner of the page. In fact, it can also have another form of expression, just like Baidu’s search records. How to implement it depends on the actual usage scenario.
5. Design the search interface
After completing the above things, you can now start designing the search interface. When designing the search interface, the following points are also included:
search bar. A common one is a search box, where users can directly search for what they want. Then there will also be some special scenarios that require multiple search boxes, such as buying air tickets, train tickets, etc. At least three search boxes for origin, destination and date are required. So when designing a search box, unless you really need multiple search boxes, it’s best to have only one search box.
Autocomplete and autosuggest. This is more common when we buy a ticket. For example, if I enter North, all lists related to North will appear. Beijing, Beihai, etc. can let us quickly choose. Another is in the search website, for example, when we enter W, content related to w will appear immediately, such as we, windows, etc.
Advanced Search. Usually only used in more professional websites. Commonly used are Boolean operations, including and, or, not.
Modification is supported. After we search, we find that the search results do not seem to match the content we want to find, and we want to modify the search content. This is very common in our life, and we can search at any time and modify it at any time. Also, when there are many search results, such as the search results of Douban just mentioned, then we should be able to let users decide to expand or narrow the search scope.
In another case, when the searched information is time-sensitive, we should consider filtering by time. Most of the product designs we use every day integrate search and browsing. Some browse after search, and some search after browse. The former is the browsing result after the normal search is completed. The latter is to filter, sort, etc. when browsing.
when the user is stuck. This kind of problem generally occurs in two situations. One is that there are too many results, and the user suddenly does not know what to do. Another is that there is no result. For the former, we can add filtering methods to allow users to conduct precise searches. For the latter, we can give suggestions to optimize the search or recommend related results.
7. Thesaurus, controlled words and metadata
The content of this part is rare for users, because they are all back-end content. This part briefly talks about metadata, controlled vocabulary, technical terms, thesaurus types, semantic relations, multi-level structure, and faceted taxonomy.
Metadata should be a word that product managers see more when communicating with technology. This generally refers to the attributes that describe the data. Such as gender, height, age, etc.
2. Controlled Vocabulary
The controlled vocabulary does not sound very easy to understand, if it is popular, it is actually an equivalent vocabulary. It mainly includes the following parts:
- synonym ring. A synonym ring connects a set of words defined as equivalence. Such as iPadtouch and iTouch. We all know they mean the same thing. So how do we let search engines know when we search? That is to use definitions such as synonym rings to edit controlled vocabularies;
- Specification document. Similar to our daily format for writing product documentation. Its terminology will mainly be standardized here to define its preferred terms and acceptable values. And will make a list according to the rules;
- classification scheme. This taxonomy is similar to our zip code. For example, 010 represents Beijing. After unifying the rules, an arrangement of preferred terms is carried out.
3. Technical terms
The techniques here are mainly used to convey the core terms of definitions and relationships. Mainly include the following:
- Preferred belongs to (PT). Also commonly referred to as an acceptable generic term. All relationships are defined in terms of preferred belonging;
- Heterosexual Term (VT). be defined as equivalent to, or substantially the same as, the Preferred Term;
- Generic term (BT). It is the upper-level term of the preferred term, which is at the first level of the hierarchy;
- Subclass Term (NT). It is the subordinate term of the preferred term, which is at the lowest level of the hierarchy;
- Related Terminology (RT). It is linked to the first term through an association relationship;
- use (U). Opposite-sex terms use preferred terms, generally using such syntax as a tool for indexers;
- for (UF). Indicates the interrelationship of the preferred term UF heterosexual term;
- Range Notes (SN). Primarily a specific type of preferred term that defines the meaning of the term.
The thesaurus is when you search for some content, there may be no keywords in the content, but it can be searched because the subject or description of the content contains the words you searched for. Thesaurus also includes the following types, and decisions are based on how you plan to use the thesaurus, which will have a major impact on your design.
5. Semantic relationships
Semantic relations mainly include equivalence, hierarchy, and association.
Equivalence is just mentioned, that is, ipadtouch and iTouch are equivalent.
There are three types of hierarchy, the first is belonging, that is, the child inherits the characteristics of the parent, such as magpies belong to birds; the second is whole-part, that is, the child is part of the parent, such as the finger is part of the hand; the third Species are instances, generally abstract, such as Jiang and Changjiang.
Association is that two things are related, for example the association of salt water and sugar water is water.
6. Multi-level structure
The structure we see in daily life is that each next level will have more or equal number of elements than the previous level, while for hierarchical structures, the next level will have fewer elements than the previous level.
The result of this irregular hierarchy is a multi-level structure, which is unavoidable when you are dealing with large information systems.
7. Faceted Taxonomy
The faceted taxonomy is actually to classify content from multiple aspects, usually from the aspects of subject, product, document type, user, geographic location, and price.
Regarding information architecture, it plays a pivotal role in daily product work; there are also some things that seem tangible, but are actually intangible.
The above content is only from a theoretical point of view, to give you a brief introduction. How to apply it in practical work, I believe that everyone will sum up a set of their own methodology after continuous practice based on the actual work situation.