Analyzing website performance

1. A computer-implemented method comprising:

retrieving, by a detection system from a capture agent, a plurality of timing metrics for a set of webpages of a website, wherein the capture agent is a software agent embedded to the set of webpages, wherein the plurality of timing metrics are associated with a plurality of network operations involving the set of webpages, the set of webpages comprising of a plurality of stages including a target stage;

for each webpage in the set of webpages:

determining, by the detection system, a timing distribution for the webpage based on a plurality of timing values for a timing metric from the plurality of timing metrics, wherein the plurality of timing values are associated with a plurality of web sessions on a plurality of user devices involving the webpage, and wherein the capture agent monitors and captures the plurality of timing values for the plurality of web sessions;

identifying, by the detection system, a first range of the timing distribution for a slow group of users;

identifying, by the detection system, a second range of the timing distribution for a fast group of users, wherein the second range is greater than the first range;

computing, by the detection system, a first rate of the slow group of users for the webpage reaching the target stage and a second rate of the fast group of users for the webpage accessing the target stage;

identifying, by the detection system, a number of users that reached the target stage;

determining, by the detection system, an impact classification for the webpage based on the first rate, the second rate, and the number of users that reached the target stage; and

displaying, by the detection system, the impact classification for each webpage in the set of webpages for a user to identify at least one network operation contributing to poor performance of the website.

2. The computer-implemented method of claim 1 , wherein determining the impact classification for the webpage comprises:

determining, by the detection system, a number of missed accesses to the target stage based on the first rate for the slow group of users;

computing, by the detection system, a score for the webpage based the number of missed accesses to the target stage and the number of users that reached the target stage; and

determining, by the detection system, the impact classification for the webpage based on the computed score for the webpage.

3. The computer-implemented method of claim 2 , wherein determining the number of missed accesses to the target stage comprises:

determining, by the detection system, a number of slow sessions for the slow group of users and a number of users from the slow group of users reaching the target stage;

multiplying, by the detection system, the number of slow sessions for the slow group of users and the second rate of the fast group of users; and

subtracting, by the detection system, the number of users from the slow group of users reaching the target stage.

4. The computer-implemented method of claim 2 , wherein computing the score for the webpage comprises:

dividing, by the detection system, the number of missed accesses to the target stage by the number of users for the webpage that reached the target stage.

5. The computer-implemented method of claim 2 , wherein the impact classification for the webpage is one of low, medium, or high based on the computed score of the webpage.

6. The computer-implemented method of claim 2 , further comprising:

identifying, by the detection system, at least one network operation and at least one attribute associated with the network operation impacting performance of the website based on the plurality of timing values and the computed score for the webpage.

7. The computer-implemented method of claim 6 , wherein the at least one attribute is a portion or a specific building block of the webpage impacting the performance of the website.

8. The computer-implemented method of claim 1 , wherein the timing distribution is determined based on a tri-mean median of the plurality of timing values for the webpage.

9. The computer-implemented method of claim 1 , wherein the first rate is computed based upon a number of the web sessions involving the webpage and a number of the web sessions involving the slow group of users that reached the target stage and wherein the second rate is computed based upon a number of the web sessions involving the webpage and a number of web sessions involving the fast group of users that reached the target stage.

10. The computer-implemented method of claim 1 , wherein the plurality of timing metrics are at least one of a server response time, a client response time, a webpage load time, a Document Object Model (DOM) response time, and a Single Page Application (SPA) transition time.

11. The computer-implemented method of claim 1 , wherein the target stage is a resulting webpage accessed by the slow group of users or the fast group of users to achieve a particular result or an objective, wherein the particular result is at least one of placing an order, registering for a website, signing up for a mailing list, opening a link, or performing any other action.

12. The computer-implemented method of claim 1 , wherein the capture agent includes Javascript.

13. A system comprising:

one or more processors; and

a non-transitory computer-readable medium including instructions that, when executed by the one or more processors, cause the one or more processors to:

retrieve, from a capture agent, a plurality of timing metrics for a set of webpages of a web site, wherein the capture agent is a software agent embedded to the set of webpages, wherein the plurality of timing metrics are associated with a plurality of network operations involving the set of webpages, the set of webpages comprising of a plurality of stages including a target stage;

for each webpage in the set of webpages:

determine a timing distribution for the webpage based on a plurality of timing values for a timing metric from the plurality of timing metrics, wherein the plurality of timing values are associated with a plurality of web sessions on a plurality of user devices involving the webpage, and wherein the capture agent monitors and captures the plurality of timing values for the plurality of web sessions;

identify a first range of the timing distribution for a slow group of users;

identify a second range of the timing distribution for a fast group of users, wherein the second range is greater than the first range;

compute a first rate of the slow group of users for the webpage reaching the target stage and a second rate of the fast group of users for the webpage accessing the target stage;

identify a number of users that reached the target stage;

determine an impact classification for the webpage based on the first rate, the second rate, and the number of users that reached the target stage; and

display the impact classification for each webpage in the set of webpages for a user to identify at least one network operation contributing to poor performance of the website.

14. The system of claim 13 , wherein to determine the impact classification for the webpage comprises the instructions, the one or more processors further cause the one or more processors to:

determine a number of missed accesses to the target stage based on the first rate of the slow group of users;

compute a score for the webpage based the number of missed accesses to the target stage and the number of users that reached the target stage; and

determine the impact classification for the webpage based on the computed score for the webpage.

15. The system of claim 14 , wherein to determine the number of missed accesses to the target stage, the one or more processors further cause the one or more processors to:

determine a number of slow sessions for the slow group of users and a number of users from the slow group of users reaching the target stage;

multiply the number of slow sessions for the slow group of users and the second rate of the fast group of users; and

subtract the number of users from the slow group of users reaching the target stage.

16. The system of claim 14 , wherein the impact classification for the webpage is one of low, medium, or high based on the computed score of the webpage.

17. The system of claim 13 , wherein the timing distribution is determined based on a tri-mean median of the plurality of timing values for the webpage.

18. The system of claim 13 , wherein the plurality of timing metrics are at least one of a server response time, a client response time, a webpage load time, a Document Object Model (DOM) response time, and a Single Page Application (SPA) transition time.

19. The system of claim 13 , wherein the first rate is computed based upon a number of the web sessions involving the webpage and a number of the web sessions involving the slow group of users that reached the target stage and wherein the second rate is computed based upon a number of the web sessions involving the webpage and a number of web sessions involving the fast group of users that reached the target stage.

20. The system of claim 13 , wherein the target stage is a resulting webpage accessed by the slow group of users or the fast group of users to achieve a particular result or an objective, wherein the particular result is at least one of placing an order, registering for a website, signing up for a mailing list, opening a link, or performing any other action.