What is a recommendation engine?Introducing the merits, how to choose, and three recommended selections

What is a recommendation engine?

A recommendation engine is a tool that a system determines and proposes information that is highly interesting for users of the same attribute, based on the site browsing information and purchase information of other users for site users.

For example, EC malls such as Amazon and Rakuten have a recommendation engine.

Based on the behavior history and purchase history of users who performed similar search in the past according to their search words, "People who saw this product also saw this product" and "Recommended for you."The function that introduces the product in a format such as "here is a recommendation engine.

The proposal and introduction function of the recommendation engine is mainly realized by the following technologies.

The recommendation engine can be built on your own and introduced it to the site, but if you do not have a high -programming knowledge, if you do not have expertise, you should use a recommendation engine developed by another company.Yes.

A wide variety of tools are sold, from inexpensive items that have been narrowed down to the site of the site and the advanced items to be converted to products on their own sites and content introductions.

Advantages and disadvantages of recommendation engine

Next, let's take a look at the advantages and disadvantages of introducing a recommendation engine.

There are three benefits when introducing a recommendation engine.

I will explain in detail in order.

If you introduce a recommendation engine, you can accurately introduce the products that match the user's preference, so the probability of purchasing will be improved.

For example, suppose there is a similar vacuum cleaner called vacuum cleaner A, vacuum cleaner B, vacuum cleaner C, and vacuum cleaner B is often purchased.Also, suppose you know that the user who is browsing the page of vacuum cleaner A has purchased a vacuum cleaner B as a result.

At this time, if you use a recommendation engine, a user looking at the vacuum cleaner A will say, "For those who are looking at this vacuum cleaner (vacuum cleaner A), this vacuum cleaner (vacuum cleaner B).I recommend it. "

On the EC site, looking at only specific products, I did not notice the existence of other similar products, and said, "I definitely want a vacuum cleaner, but I don't really want it."There are many cases that cannot be reached.

Therefore, you can appeal that you have a vacuum cleaner that you really want without missing the opportunity, so you can expect an increase in the probability of purchasing.

If you introduce a recommendation engine, it is also an advantage that you can propose a product B and a product C that is easy to consider together with users who are trying to purchase Product A and encourage additional purchases.

For example, from the data of the past buyers, it is supposed that users who purchase a comforter tend to buy a comforter cover at the same time.

At this time, in that case, using a recommendation engine, in that case, for users who are browsing the product page of the comforter, "For those who are looking at this product (comforter), this product (comforter cover) is also available.I recommend it. "

Even in the apparel company I was in charge of, I introduced a recommendation engine and displayed the bottoms for coordination to those who are looking at the tops, but the unit price of the customer was 120 % year -on -year.

By introducing products and content that suits users's preferences, it will be a highly convenient site for users, and can prevent withdrawal in a short time from visiting the site.

The client in charge of the client operated an EC site for the accessory business, but when the recommendation engine was introduced to enhance the migration of the site, the site's migration became 200%.

As a result, the time of the site has increased, and the conversion rate has improved and sales have increased.

On the other hand, the recommendation engine has the following disadvantages, so be careful when using it.

If the number of products and content are small, the number of recommendations is small, making it difficult to introduce products and content according to the user's preference.

The promotion of users and the convenience of the user will not be as expected.

For example, if there is a site A with 10 items and a site B of 1,000 items, it is clear which site is noticeable the effect of the recommendation engine.

In the industry where many products and abundant variations are essential, you should prepare 100 products if possible.

Although it is not easy to say how much the number of products and the number of content is necessary, there are cases where more than 200 items are prepared on the stationary goods EC site, and the effect of the recommendation engine has been noticeable.

If there are few users, the accuracy of the recommendation may not increase, and there may not be a recommendation that suits the user's preference, which may not lead to promotion of purchase and convenience.

Therefore, if you have just launched an EC site or a website, and there are still few visitors, it is better to predict the user's behavior pattern and set the contents of the recommendation manually.

How to choose a recommendation engine

The recommendation engine offers a wide variety of tools, but if you select a tool that suits your company, we recommend that you value the following three points.

Be sure to check before you sign up for the compatibility of the site you are currently using and the recommendation engine.

For example, depending on the online shop system you are using, there may be limited recommendations that can be introduced.

When checking compatibility, please tell the system vendor the online shop system and CMS you are currently using and check for the presence.

The functions that can be done with a recommendation engine are different depending on the tool.

"Tools that record based on the user's sites and purchase information on the site", "Tools that take into account the user's site and incorporate the hobby taste outside the user's site" "manually optional".There are many functions that can be done by the recommendation engine, such as "tools that recommend information.

Please refer to the necessary functions for each common purpose (used example) by introducing the recommendation engine.

The number of settings and operations during the introduction of the recommendation engine is also different depending on the tool.

For example, depending on the online shop system you are using, some tools can start in a few minutes to tens of minutes as long as you have a contract, while others cannot start without complicated rules.

In addition, some things can be operated by automatic settings after the introduction, while others will not be effective unless the settings are changed in detail.

The tools that need to be set in detail can be said to be a tool with a high degree of freedom if you turn it over, but let's calmly determine if your operational system and man -hours are worth it.

When making a decision, we recommend that you check the system vendor for the average work time required for introduction and the subsequent maintenance time.

Recommended recommendation engine

Finally, here are three recommended recommendations.

EC Recommender is a recommendation engine that can be started at a low price according to what you want to do because the plan is divided depending on the function.

The cheapest plan includes text mining functions that propose products that are highly related based on text data of product information, and a recommendation function based on browsing information from others.

Therefore, it is a good idea to use the plan first and change it to a higher plan when you gradually understand the necessary functions.

Sabumi!Recomend is a record engine from a large industry, with a track record of introduction in a wide variety of industries and business formats.

RTOASTER is a recommendation engine that incorporates a wide variety of data in addition to the behavior data on the site and realizes a highly accurate record that has been personalized by each person.

At the end

Now is the time when information is saturated.Users need to choose information suitable for themselves from a lot of information, which is not easy.

The recommendation engine is a tool that can read the behavior pattern of users and provide suitable information for each user.

The provision of services that faces each user will always lead to improving the satisfaction of the user, so use the recommendation engine to make the most of your service.