Loyal customers can not only create lasting value for the site is also the site an important channel for brand reputation to promote, so the current site of loyal users pay more attention to. The user may be many sites or web analytics tool division of the "New User" and "returning visitors", but simply to distinguish between old and new users is not enough, we need a more comprehensive indicator to measure the loyalty of users of the Site.
User loyalty (the Loyalty), refers to the user for the extent of repeat purchase of the business or brand preferences and recurrent. For web sites, customer loyalty is out of site features or preferences frequently visit the site. According to the theory of customer loyalty, loyalty can be measured by the following four indicators:
- Repeat purchase intention (Repurchase Intention): willingness to purchase the type of product purchased;
- Cross-buying intention (Cross-buying Intention): to purchase previously for the type of product purchased or the willingness to expand their services;
- Recommendation intention (the Customer Reference Intention): recommendation to other potential customers pass the wishes of the brand reputation;
- Price endurance (Price Tolerance): customers are willing to pay the highest price.
Quantify the site's customer loyalty
The above four indicators for e-commerce sites, there may be applicability, but for most sites is not appropriate, so in order to allow analysis of general applicability, in order to meet all the indicators can be quantified (see above intention to recommend more difficult to quantify), for quantitative analysis of the requirements, where you can select Google Analytics customer loyalty four metrics: Repeated Times, Recency, Length of Visit, Depth of Visit, the frequency of user access, the recent access time, the average residence time, with an average visit to a number of pages, these indicators can be calculated directly from the Web site click-stream data, are applicable to all sites, the following look at the definition of these indicators and how to calculate the (site metrics The definition of reference - web analytics metrics ):
- Frequency of access: users visit the site the number of times within a period of time, each user the number of Visits;
- The last access time: users visit the site time, because this indicator is the concept of a point in time, so easy to measure, and generally the user access time from the current number of days.
- The average residence time: the average residence time of each visit the user for some time, namely the number of each user Time on the Site's and / Visits;
- The average visit to the page: the user for some time the average per visit browsing the page number, ie, the number of each user. Page Views / Visits.
Time interval of the statistical data is based on the characteristics of the site, if site information updates faster, more frequent users to access, you can properly select a shorter time, such data changes on the sensitivity will be higher; the contrary, the choose a slightly longer time period, so that the user's data is richer, and the results of the analysis of indicators will be more accurate and effective.
Display of customer loyalty and comparison
Above four indicators can be quantified statistics to get a single indicator does not make sense, we need to find out what is a loyal user by comparing, which is the loss of some processing can be indicators, so that between them more comparable, you can refer to the previous article - the standardization of data ), I used the standardized methods of min-max, first of all index values converted to the interval [0,1], then the magnification such as 10-point scale to rate, by 10, the data are all distributed in the interval [0,10], as shown below:
| Frequency of visits | Last access time | Mean residence time | Average access the page | ||
| User 1 | Data | 2 | 15 days ago | 150 seconds | 3 |
| Standardization | 0.10 | 0.50 | 0.30 | 0.38 | |
| Score | A | 5 | 3 | 3.8 | |
| User 2 | Data | 8 | 2 days ago | 120 seconds | 5 |
| Standardization | 0.40 | 0.93 | 0.24 | 0.63 | |
| Score | 4 | 9.3 | 2.4 | 6.3 |
- Data in the table just a simple example, the actual situation of the minimum and maximum values of each indicator is calculated
According to the data of the above table, we have all the indicators into line with the same score range, then you can use the radar chart to display the user loyalty. Radar chart display has the following advantages:
- Show all evaluation;
- The user bias in the scores of each indicator;
- A simple analysis of customer loyalty score, graphics surrounded the area (assuming that the four indicators equal weight and importance of the existence of a significant difference, the area can not be used to measure);
- Can be used for the comparison of user loyalty.
The following is a sample radar chart drawn on the basis of the above table:
The meaning of customer loyalty
Then based on the results of this show, what can we do it? In fact, for any site, there are two directions is the same: to retain loyal customers, reducing the loss of users. Customer loyalty based on the above evaluation system expanded to is:
- Analysis of the behavioral characteristics of loyal customers, and strive to meet their needs, enhance their satisfaction;
- From the last access time indicators modem customer loyalty trends may be the loss of users, to analyze their possible reasons for the loss of, and attempt to retain the loss of the user;
- Loyal users and the loss of users of differences in the values of the indicators, to find which indicators the gap has led to the reduction of customer loyalty, and optimize site performance in these areas.
So, I am using here is based on user access frequency visit time average residence time, the average access number of pages of these four indicators to evaluate the site's users loyalty, and radar chart carried out show and compare, maybe you can based on your own site characteristics to find a more suitable indicators and the presentation, and ultimately need to do is to more accurately locate the site of loyal users, and efforts to retain them.
»In this paper, the BY-NC-SA agreement, reproduced please specify source: The data analysis » Website user loyalty analysis
Related Articles:
Hi, very good.
But the feeling of your analysis more engagement rather than loyalty. As you said, the Loyalty for existing customers to talk about.
Analysis of the engagement, then, Eric T. Peterson, a more detailed analysis, you can refer to:
http://blog.webanalyticsdemystified.com/weblog/2007/01/engagement-metric-defined-part-iv-in.html
@ days shore : Thank you to the recommended articles. the indicators of engagement may be different, depending on the site, I just want to find a way for all site users to analyze the text of the four indicators can be used as a basic measure to do the extended analysis, or analysis can also be called user viscosity analysis, only the "loyalty" I borrow the concept of the GA.
Can the above user and user data standardization process issued I do not thank you
The the _AT_ lesk : Hello, the article's data is only my analog data. You should be no problem in the standardization of data min-max standardized methods for data processing.
Where max is the maximum value of the sample data set for each indicator, min is the minimum value, x is the sample index value calculated, if the amount of data is not great with Excel can handle.
Have more questions, you can add my Gtalk: joeghwu@gmail.com
Data assume that max = 20 min = 0 pairs.
Zhaoruxin : Hello, Min and Max can be obtained through the minimum and maximum values of the actual data, without assumption.
Will the last access time is the inverse indicator?
_AT_ nancy : Well, so to speak. Because a recent visit to distance the closer the user the higher the viscosity, the farther from the current point in time, potential users are moving away from your site.
I prefer to use a recent visit to the interval (in days) to replace the last access time of this indicator; simple point in time into a quantitative measure.
Hello, I have some questions about graphics surrounded the area of the radar chart as a customer loyalty score (personal opinion): First of all, if the indicators of changes in the order (in fact, the radar chart in the article there are three painting), graphics surrounded the area of the radar chart is capricious, the radar chart area is not unique; In addition, I made reference to some information, some with graphical radar chart enclosed area multiplied by the corresponding graphical and then prescribing the perimeter, area multiplied by the side length of the square and then prescribing as a composite score.
_AT_ bobo : Well, I mean here is a simple comparison, not from the geometric meaning of the radar map of the area surrounded necessarily represent a composite score, if you need a more accurate composite score can be considered after the indicators standardized weighted .
hi, ask a question:
Why should the user loyalty is divided into four dimensions I want to know these four dimensions are necessary and sufficient?
I think a good criteria for the classification should be complete and independent
A frequency of visits, the last access time of these two values a little overlap, if the user access to the flat rate is the week, the average access time is 3.5 days
2 before the mention of "cross-purchase intention and recommend intention," not from the indicators reflect the hope to discuss, thank you
@ root : Thank you very question you raised.
Customer loyalty is not a directly quantifiable indicators, we need to be reflected from the other statistics indicators; example, article 4 indicators are not sufficient nor is it necessary, this example is not a site analysis the standard model of customer loyalty. site analyst according to the operational characteristics of the site to build loyalty analysis model, indicators are not limited to these four, but the four indicators are almost all web sites can be, so take a little reference, a good analytical model is The key to an effective conclusion.
The definition of "access frequency" Please look, it is not access to the average interval, but the number of visits within the period of time, so there is a difference with the last access time.
Article mentioned at the beginning "cross purchase" recommendation intention "is the traditional definition of customer loyalty, the article level, web-based analysis in order to get quantitative indicators selected user access to the site, and purchasing behavior in e-commerce can be analyzed in the same site, you can look at another of my articles.
Joegh : Is there no good electricity supplier website loyalty analysis model? Quantitative measurement that I intend to do a loyalty evaluation system, but do not have the model, model papers of previous years, mostly through questionnaires. Thank you
_AT_ coder : I think if you want to analyze loyalty, not necessarily readily available quantitative models, not many may indeed be entirely quantitative. But should have a lot of elaborate loyalty thesis, in fact, from a discussion of some of the loyalty to the establishment of a viable analytical model is also a questionnaire As for the data acquisition, the actual data to full access to the site may indeed have a certain difficulty, but also to consider the validity of conclusions based on these data.