System and method for automated creation of a load test plan

1. A computer-implemented method for creation of a test plan for load testing a website or web application comprising:

receiving a set of input parameters, which includes a specified time period range;

retrieving, from a data storage repository, beacon data of real user sessions on the website or web application that fall within the specified time period range;

identifying, from the beacon data, a set of peak traffic time periods, which includes a peak day, a peak hour, a peak minute, and a peak second;

calculating user session statistics for the peak day, the user session statistics including median session duration, session length, and think times;

generating a page group distribution table that includes a list of all page groups and page group hit percentages, relative to all other page groups, for the real user sessions during the peak hour; and

generating a set of test scenario paths representative of the page group distribution table.

2. The computer-implemented method of claim 1 wherein the generating of the set of test scenario paths comprises:

a) assigning weights to each page group in accordance with the page group hit percentage of the page group;

b) creating a list of all the real user sessions during the peak hour having a session length greater than 1, for each real user session in the list determining:

(i) an actual path, which is a sequence of page groups traversed by the real user;

(ii) a page group list which includes distinct page groups in the actual path;

(iii) a total page group weight, which is a sum of all the page group weights for the actual path;

(iv) a session length deviation, which is an absolute value of a difference between a session length of the real user session and the median session length;

(v) a session duration deviation, which is an absolute value of a difference between a session duration of the real user session and the median session duration;

(vi) an exit-entry proportion, which is a proportion of all the real user sessions that have the same entry and exit pages;

(vii) an exit-entry weight, which is a number of real user sessions that have the same entry and exit pages;

(viii) an exit-entry max paths, which is calculated by multiplying the exit-entry proportion by a specified number that limits the maximum paths;

(ix) a primary sort variable, which is a ratio of the session length deviation to the total page group weight; and

(x) a secondary sort variable, which is a ratio of the session duration deviation to the exit-entry weight;

c) sorting the list of all the real user sessions during the peak hour by the primary sort variable, and then by the secondary sort variable;

d) assigning a representative rank to each session according to position in the sorted list;

e) determining a number of critical page groups by sorting the list of all page groups in descending order according to the page group hit percentage of the page group, the number of critical page groups being equal to a number of page groups from the sorted list of all page groups where a cumulative page group hit percentage is less than or equal to a minimum specified page coverage;

f) selecting a top n unconverted session paths from the sorted list of all the real user sessions, where n is a specified maximum number of abandonment paths, top n unconverted session paths comprising abandonment paths of the set of test scenario paths; and

g) selecting a top n converted session paths from the sorted list of all the real user sessions, where n is a specified maximum number of abandonment paths, top n unconverted session paths comprising conversion paths of the set of test scenario paths.

3. The computer-implemented method of claim 1 further comprising creating a master table that lists the set of load test scenarios and the think times that represents a distribution of paths taken by real users during the peak hour.

4. The computer-implemented method of claim 2 wherein the selecting of the top n unconverted session paths further comprises only selecting unconverted sessions that meet a filtering criteria.

5. The computer-implemented method of claim 4 wherein the filtering criteria comprises a number of page groups in the page group list is greater than 1, and the representative rank is less than or equal to the exit-entry max paths.

6. The computer-implemented method of claim 2 wherein the selecting of the top n converted session paths further comprises only selecting converted sessions that meet a filtering criteria.

7. The computer-implemented method of claim 6 wherein the filtering criteria comprises a number of page groups in the page group list is greater than 1, and the representative rank is less than or equal to the exit-entry max paths.

8. The computer-implemented method of claim 1 where in the generating a set of test scenario paths comprises determining one or more bounce paths for inclusion in the set of test scenario paths.

9. The computer-implemented method of claim 8 wherein the determining of the one or more bounce paths comprises:

generating a list of distinct landing pages for the real user sessions that have a session length equal to 1;

determining a bounce path percentage for each distinct landing page in the list, the bounce path percentage being calculated as a percentage of sessions that bounced on the distinct landing page relative to other landing pages in the list;

ranking the list of distinct landing pages according to the bounce path percentage, from highest to lowest;

calculating a cumulative bounce path percentage, starting from a highest bounce path percentage, for each distinct landing page in the ranked list; and

selecting as the one or more bounce paths for inclusion in the set of test scenario paths, those distinct landing pages where the cumulative bounce path percentage is less than or equal to a specified minimum bounce path coverage.