Editor's note: the long tail effect is not a strange word. The long tail effect was originally published in Wired magazine in 2004 by Chris Anderson, editor in chief of Wired magazine. This word reflects the "neglected part" of websites, products and business models, which is often neglected in small quantity, many categories, but a huge total. This phenomenon is particularly obvious in the field of Internet and digital products, and it is also in the "neglected" part of the "Pareto principle" discussed in the previous article. Today's article, also from the famous nngroup, discusses in detail how to find new opportunities with the help of the long tail effect in the field of digital products.
For a long time, the channels for people to obtain information in life are relatively fixed. You need to go to the brand store to browse the new clothes styles and to the bookstore to check the newly arrived books. The way you get to know a new band is usually to listen to the introduction of the clerk when you go to a concert, buy a tape or CD, or see it on the poster, or you will have one or two friends with good taste or minority music taste to listen to their introduction. In fact, the ways and channels for people to understand information are diverse on the whole, but relatively fixed. These can also be well reflected in the sales data.
When you look at sales data, you will find that they usually conform to power-law distribution. In other words, a small number of dominant head products obviously account for the majority, while a large number of low sales products account for the long tail.
The definition of long tail usually refers to the total amount of data close to the tail according to the power-law distribution. The charm of the long tail effect is that the actual accumulated total amount of these products or projects with few sales actually represents extremely huge comprehensive income.
Over time, the items or products generally located at the head will change. With the increase of new content, the content and commodities of the long tail will also increase and extend, and this extension will usually make the total amount of these commodities or contents with less sales occupy a larger share and become more important.
Before the arrival of the digital revolution, because the acquisition, sales and distribution of goods are restricted by strict physical and economic conditions, enterprises need to accurately predict which goods are better sold as much as possible before everything starts, so as to urge enterprises to aim at a broader market and hope to ensure the overall sales volume by helping goods closer to the head and higher weight.
Popular goods will become popular more easily with the positive feedback cycle (that is, the "success breeds success" effect), and the birth of popular goods will correspondingly push more other goods to the position of the long tail. In fact, social media is very similar. Celebrities will have more and better contacts, and they will be easier to accumulate contacts and obtain resources faster than most users, which is due to the iteration of existing contacts and resources.
After 2000, the e-commerce revolution appeared. Wired magazine reporter Chris Anderson observed that digital platforms such as Amazon, Netflix and iTunes are more likely to benefit from the long tail... Because they are less vulnerable to physical constraints, He derived such a conclusion from a series of cases: online channels can push more obscure and niche content and products in the long tail to global users. The strategy of making full use of the long tail means that although the number of users covered by a single commodity and project is small, the total amount of such commodities is very large, and it is very feasible to provide services for a wide range of user groups.
This long tail effect is more obvious in the field of Internet and digital products:
- Generate more revenue in e-commerce and social media
- Drive traffic growth and reduce marketing costs
- Enhance the overall influence of the website and app
- Improve the accuracy of machine learning
Increase revenue from e-commerce and social media
In the decades after Anderson put forward the long tail effect, logistics, search, network hosting, demand forecasting, e-commerce and other fields have steadily increased and innovated. The improvements in these areas have led many enterprises to gradually shift their attention to popular products to the creation of diversity, customization and content. Many websites and apps allow users to customize the product itself. Social media platforms such as youtube and instagram rely on users to contribute long tail content, while online shopping platforms such as Amazon and tmall collect millions of different categories of goods through powerful search and filtering functions.
However, the long tail strategy only applies to e-commerce and social media:
- Help users successfully find, customize and buy relatively small needs and goods
- Help users participate in content creation more quickly and successfully
However, in order to support these requirements, the UX team must monitor the entire information architecture and search results, do a good job in user guidance, understand various possibilities of custom workflow, cooperate with the product department, adjust the product roadmap, and strive to integrate these functions into the product core, not just for the implementation of new functions.
Key points for user experience design to better create revenue:
- Enhance the searchability and discoverability of content, and improve the information architecture and navigation based on it
- Improve the availability of search functions to help users filter to specific products that are more suitable for them
- Conduct qualitative usability test on product customization workflow to ensure that users can customize more unique and targeted products
- Benchmarking the content creation process with quantitative availability analysis to monitor efficiency
- Help users fit the overall user experience faster through context help, so as to simplify user guidance
- Use "e-commerce best practices" to check the existing design. According to today's standards, the vast majority of e-commerce workflow can be improved
Increase traffic and reduce marketing costs
The vast majority of web search results are brought by a disproportionate number of popular keywords.
The long tail effect also has a great influence in search engine optimization (SEO) and search engine sound (SEM). Indeed, some keywords will get a large number of queries disproportionately, while others form the long tail of website search results. Unfortunately, the fierce competition will lead to an extremely high advertising cost of popular keywords, and the use of multiple long tail keywords can effectively reduce competition, reduce advertising costs, and meet the specific needs of users.
UX and marketing teams can work together to tap the value of long tail keywords to attract and retain users. By studying user habits and personalized needs, and cooperating with the marketing department, we can provide users with customized content and control costs more accurately.
- The long tail keywords of users are obtained by analyzing the search log of the site
- Through qualitative research on users, this paper reveals how enterprises conduct research on users with specific interests
- Use better content strategies to fit the search engine, so as to obtain higher ranking
Reduce the impact of website delays
Most page requests are directly related to page loading time. When the traffic of the whole website is very large, relatively slow long tail requests will increase user loss and reduce revenue.
The delay of the website will also show a long tail mode. Although the response time of the website is directly related to the number of requests, the long tail accumulated by pages with poor performance will turn into a large number of user losses and directly affect the revenue of the website.
It is recommended that the design team work with the IT department to give priority to reducing the overall delay. Improving the overall responsiveness of the website through server optimization or server upgrade may bring more benefits than experimental design schemes.
Strategies to solve delayed UX promotion:
- With the help of the website analysis tool, check the response speed of the website and notice the abnormal value of the rate in the page loading
- Monitor users' information feedback channels and collect and sort out reports on various problems. Remember, feedback doesn't mean only these problems, because a large number of users will choose to leave directly instead of feedback when facing a bad experience.
Reducing the error rate of the algorithm requires both common examples and rare cases, and the training set composed of the two should be as many as possible.
Long tail data plays an important role in the training set of machine learning. Machine learning needs a lot of data to train AI, and build a more accurate model based on these data to predict the future. In the early stage of the project, the company can use the public data set to train AI, which is easy to obtain and economical enough. However, with the improvement of accuracy requirements, if you want to continue to reduce the probability of AI errors, it involves more and more rare and unusual cases, which belong to the long tail data.
Inaccurate AI will gradually be rejected by users. In this case, the importance of long tail data is self-evident. However, obtaining long tail data is an important challenge. At this time, UX designers need to cooperate with colleagues in the data science department to jointly deploy and design a high-quality test environment - minimize the poor state of users when providing training data, make users trust, not affected by irrelevant factors, and focus more on providing data to help AI improve, rather than being affected by aggressive data collection needs. In fact, designers need to remind data experts that the core demand of users is to achieve goals, not to train models for them - non-invasive data collection is the key.
- At this time, the value of user experience design is mainly reflected in:
- Follow the best practices of the recommendation system to promote the discovery of long tail content and user participation of the platform
- Over time, build trust with users as much as possible to improve the accuracy of data collection in an appropriate way. In the guidance stage of new users, try not to ask for personal data.
- When the accuracy of machine learning cannot be done well enough due to insufficient data, try to maintain a sense of trust with users with the help of design and copywriting.
- Participate in qualitative research to understand users' views on the practicability of machine learning.
- By reducing the interaction cost of users submitting feedback to the system, we can get effective user feedback faster and better.
- When users complete the task, they can judge through better identification mechanism and track user behavior to enhance implicit data collection.
Is there a contradiction between Pareto principle and long tail effect?
The Pareto principle is discussed in sufficient detail in this article. Generally speaking, the Pareto principle focuses on 80% of the results rather than an additional 20%, mainly because its use scenario is a time of tight time and scarce resources.
These two concepts and functions do not conflict at all. They can work together to formulate short-term and long-term strategies. The main difference is that the two can help you screen out the different types of projects you need to focus on, whether they are a few that affect the overall situation or a large number of long tail parts. In contrast, a large number of long tail small projects can indeed bring enough impact, but the individual is not so important, so they need to be treated as a whole.
conclusion
The long tail theory is an indispensable part and strategic opportunity for business, marketing, implementation and machine learning artificial intelligence. However, it is worth noting that a single designer and UX team cannot deal with the long tail problem alone. We need to seize the opportunity and cooperate with other departments to produce an economic and highly scalable solution.
Extended reading:
