In this way, it is difficult to change the user.

The traffic comes in, if there is no conversion, the operator’s drawn work may be considered meaningless.

Transformation has always been a topic that operators concerned. How to develop an operational strategy for lifting conversion rates, inevitably inseparable from new discoveries and new insights after data analysis.

This article will introduce the basic logic and processes of transformation analysis, and share common conversion analysis methods and precautions, hoping to help operators quickly find ways to effectively improve conversion rate.

1 first point

ENGAGEMENT LOOP model product objectives

Before the conversion analysis, our heart must have a clear product goal, which will not be blindly analyzed. Product goals need to be closely related to business, so that follow-up analysis is targeted, and can help operators guide business practices. So let’s talk about how to set a reasonable product goal.

1.1 Knowing the ENGAGEMENT LOOP model

There is no target analysis, even if the problem is discovered, it is difficult to really help the product growth. Therefore, clarify the product goal is the premise of all analysis.

The following figure shows the engagement loop model, which is divided into five interconnected and cyclic phases according to the extent of user experience products. With this model, we can unclear your own business and determine the business goals and practice directions in each stage.

The five phases in the Engagement Loop model, whether it is C-terminal products or B-end products, is basically applicable:

  • The first phase, Curiosity (cause interest)

In this stage, users have a preliminary understanding of the product to produce certain interests. Through old users share, advertising, etc., if you can successfully wake up the user’s interest, you will reach the first phase of the first phase of the user to drain the user.

If a user successfully downloads or registers the product in the floor page, and completed the first start or experience, then he translates to the second phase of the participating loop.

  • The second stage, A-ha monent (merit time)

This word was originally a pleasant experience in thinking about thinking in psychology, which is equivalent to the "Mourning Open" we often say. In the product design, it refers to the moment when the user knows the product value at the beginning of the product. The product needs to bring a good experience to the user, and strive to pass your core ideas and value in a short time.

In product design, the shorter time interval from Curiosity to the A-HA Monent transformation is, the better. Because the entire transformation process is like downloading a large-capacity file, the user does not want to be caught in the middle. Help users recognize the "coolness" of the product as soon as possible, for subsequent transformation and rescue help very much.

However, different types of products require specific analysis, and the time conversion to the A-HA Monent cannot be generalized. Take the easy-to-see example as an example, the entire transformation path of B-terminal products is relatively long, and the user is initially understood from the official website, selecting the white paper or public class registered to download and interested interest, only the first phase. Later, the user experienced Demo in the official website, and then applied the product to its own business, creating an analysis and optimizing the operation process using the analysis results, and take a long time to experience the core value of the product.

But compare some social or e-commerce C-terminal products, the transformation of A-HA Monent is relatively faster. Users have added a few friends chat interaction, or in the App home page browse to the favorite item, often understand the interests of the product’s core functions.

  • The third phase, Value Exchange (Value Exchange)

After the value of the user’s approved product, we hope that the user has a certain cost exchange with us, which can be money (for example, the user is in the e-commerce platform), or it can be time, attention, data, etc.

  • The fourth stage, TRIGGER (trigger mechanism)

In order to realize the value exchange of the third phase, we tend to design many so-called triggers in the product, such as membership level, continuous sign-in reward; simultaneous manufacture, such as price cuts, distribution coupons.

The third stage and the fourth stage are the process of reciprocating, we need a complete operation action to increase the user’s viscosity of the product and precipitate the user.

  • The fifth stage, Social Proof (social identity)

Can you get the user’s universal agreement and recommendation, is the standard for testing whether the product is clearance overall. Especially in today’s highly information society, search and see other users have a very convenient evaluation of the product, and the social attributes of the product become more important.

Therefore, the setting of the target should be a stepped, different steps of the target, can correspond to the different phases of the user. Although the method of setting the goal is thousands of, the principle of setting the target is common:

  1. Confirm what the value is provided;

  2. Since the top and then set, first determine the goals you have to achieve, then extend down;

  3. The number of users who monitor each ladder target increases the ladder when the gap is too large.

  4. The ladder target is not updated regularly.

1.2 Twitter’s ladder transformation goal

Take Twitter as an example to see how the ladder target in its conversion analysis is established. The picture below shows the early days of 2010, and its product team has a seven step transformation goals set by users based on the extent to which the TWITTER uses Twitter, which is from a novice user to grow into a mature user.

In the first phase, the product has roughly understanding and impressive, the second, three-stage allows the user to experience Twitter as a social media core browsing content, if the user has fun in the browsing content, experience the so-called A-HA Monent The follow-up will enter the value exchange phase.

The user spends time and traffic, downloading the app to the mobile phone to exchange a more comfortable experience. Next, users may experience a variety of functions in Twitter, participate in content production, and even establish a social contact with other users.

As mentioned above, the basic principle of setting the ladder target, a good ladder goal must be updated in a timely manner according to the actual situation. Twitter has developed to the present, and its ladder goals have also changed a lot. Most importantly, the goal should be matched to the current product situation, clear and operability, and can guide subsequent transformation.

After obtaining such a ladder goal, subsequent transformation analysis is to serve the goal.


Complete conversion analysis process

In terms of transformation analysis, we need to monitor data between steps. From any ladder, it can be used for conversion analysis vertical or laterally.

The following is a transformation of A-HA Monent’s stage to the example, and explain how to conduct a complete conversion analysis process.

2.1 Conversion analysis of vertical direction

  • Conversion analysis node control

The conversion analysis of the longitudinal direction requires understanding and controlling the overall business conversion. Before analysis, all nodes to the A-HA Monent are needed, and the conversion funnel is created according to the node to determine the conversion cycle.

Take a look at the key node of the product to the A-HA Monent with easy view. (Note: The following data is virtual)

The first step, the user visits the official website of easy viewing boat; the second step, the user is completed in the official website, the form provided by the registration can get the basic information of the user authorization; the third step, the user actively integrated the SDK of the easy-to-seek Ark (Note: SDK is Software Development Kit, Software Development Kit), enables easy-to-see a boat to collect data to help users conduct several analysis; fourth steps, use easy view to customize the creation analysis model.

The selected node is relatively important in the conversion path core events. After the determination node, then select a suitable conversion cycle. Conversion is a conversion funnel period duration interval from the first step to the last step experienced.

In the above case, if the conversion period to seven days, while the user from accessing the official website to create an analysis after 10 days, this is actually not successful conversion, because the conversion period has exceeded long.

Face different products, we need to plan the appropriate conversion path, find a suitable period, the resulting data interpretation it may be accurate. On the basis of this step on, in order to more deeply understand the business, we will continue to increase in dimensions, broken down to work.

  • Conversion analysis is further broken down

In order to be able to analyze the conversion instructional strategies to optimize operations, we will increase the dimensions in the conversion analysis, detailed look at the case of the conversion of certain types or features. The figure below shows comparison of telephone communication as a conversion effect variables, we will find a telephone communication conversion rate will be higher.

According to this line of thinking, we will be able to be broken down by user demographics region, gender, age, etc., will be able to see a difference between conversions for different users. We should make every effort to ensure that users want services product design, overlap with the fastest population transformation.

Taking a white-collar crowd for social networking products, for example, after creating analysis by Analysys found the Ark, because some users accustomed to the operation and consistent product design, in a short time was a good experience, the conversion rate is high. Further analysis found that white-collar workers in the label of the overall user population in the conversion rate is not ideal, but a higher conversion rate of student users.

At this time there is a problem need to be adjusted, may be wrong in the design plan of the core values ??of product features, price of the product positioning, product concept or transfer mode.

In addition segmented by demographic characteristics, we can also through a variety of segments dimensions, such as the impact telephone communication, advertising and other marketing methods for the conversion of the user.

Thus, you can find the best way to improve the conversion of the target user group assigned to this channel for more budget. This is our vertical dismantling accordance with the objectives and breakdown of the purpose is to find the essence of the problems that affect conversion rates.

2.2 Analysis of lateral conversion

Lateral conversion conversion path analysis means in the longitudinal direction, we need to expand the key for each node.

In the transformation analysis, we need to look at the overall conversion data to see the effect of the conversion between the step goal. If you find that at this time the data is not ideal, we need to lateral spread for user action before reaching a core event, dismantling the lateral segment funnel.

Also at Analysys Ark, for example, the second step means that the user successfully registered at Analysys Fangzhou Guan network. However, registration is successful but the final conversion event, before the successful registration, users also need to jump a few actions.

FIG term Notes:

1.CTR (Click-Through-Rate, ie the click-through rate, refers to the network ad clicks divided by the actual amount of the ad to show)

2.LP (Landing Page, landing page)

3.PV (Page view, ie page views)

Enter from either channel user Analysys Fangzhou Guan network will first access channel from landing page, landing page contains links to registration page; then, the user fill out the form in the registration page; then, registration will be divided into several steps, For example to set a password, fill codes and the like.

Analysis transverse conversion loss for the user in a longitudinal flow. Longitudinal found churn successfully registered this critical node of serious, only 2000 successfully registered 10,000 users, conversion is only 20%, so the success of cross-sectional analysis when analyzing the registration, to tease out the refinement when users register step.

By the table below, we can see the user access path from the exposure to click to landing pages and registration pages for centralized analysis.

Assuming that search engine advertising exposure generated by the 10,000 users are served from Baidu, traffic access has successfully jumped 30%, indicating that there are only 7000 users into subsequent registration.

7000 users only 5000 visit to registration page and complete the first step before the last successful registration, the user every step of the loss. So, every step of the process can be optimized in both points, which is transforming the business value of the final analysis can bring.

Vertical and horizontal transformation analysis, corresponding to a funnel users to build layers, each layer is associated funnel business goals or a small target.

3 Third Point

Attribution analysis of transformation

The reality is often more complicated than the hypothesis, the results will be as good as the assumptions of the ideal, user behavior in the course of the visit it is diverse, so we must consider the good attribution problem, making the analysis more sophisticated and rigorous transformation.

As before users access Analysys Fangzhou Guan network, pre-produced brand exposure from all sources, to measure how effective channels to relate to the selection attribution models. Common attribution models have shown in the following figure 5 types, can be attributed to read about, "said attribution door" metaphysics ", from start to finish to find out? "

In electronic business platform, the necessity of attribution models are even more evident. For example, the user enters the Jingdong a product landing page through Baidu search, before the user has been added to the product to access the micro-channel circle of friends or other channels through the shopping cart. Users eventually completed the purchase of the product, but this step is only a Baidu search channels assists, did not directly lead to a conversion, but this step is very important.

Therefore, we should abandon the past blindly select the last contact attribution model of attribution approach, return to the traffic sources and channels of delivery itself, choosing the best attribution model, so that subsequent conversion of more scientific analysis and interpretation.

Recent White Paper