David Jones About the Author

David Jones is the Director of Sales Engineering and APM Evangelism for Dynatrace. He has been with Dynatrace for 10 years, and has 20 years’ experience working with web and mobile technologies from the first commercial HTML editor to the latest web delivery platforms and architectures. He has worked with scores of Fortune 500 organizations providing them the most recent industry best practices for web and mobile application delivery. Prior to Dynatrace he has worked at Gomez (Waltham), S1 Corp (Atlanta), Broadvision (Bay Area), Interleaf/Texcel (Waltham), i4i (Toronto) and SoftQuad (Toronto).

Black Friday / Cyber Monday 2014 Web and Mobile Performance Live Blog

Update November 25th, 2015 at 12:05AM

Did you end up here after searching for our Black Friday/Cyber Monday live blog in Google?  Here is this years (2015) live blog:

http://apmblog.dynatrace.com/2015/11/23/black-fridaycyber-monday-live-blog-retailers-mobile-web-performance/

 

Update December 1st, 2014 at 5:05PM

CyberMonday Wrap Up

Wrapping up Cyber Monday with a look at the Top Retailers who prepared the most for the onslaught of mobile traffic this Cyber Monday. Screen Shot 2014-12-01 at 4.52.40 PM
Sears, Costco, Office Depot, REI, Saks and NewEgg all understood that servicing their customers on mobile devices required a different set of priorities than delivering content to  tradition desktop browsers.  By using techniques like Responsive Web Design, these Retailers tailored the view of their sites to the mobile devices.  They provided mobile optimized experiences that reduced complexity, delivered targeted content, used less third parties and ultimately provided more satisfactory experiences.

Screen Shot 2014-12-01 at 4.54.08 PM

Update December 1st, 2014 at 4:40PM

Staples continues to have a tough CyberMonday.
Screen Shot 2014-12-01 at 4.33.06 PM
We all have to remember that Staples is working with a very complex application, with a lot of moving parts, but this is not different from any of the other top retailers.  They also have to manage an incredible amount of complexity.  What we know from working with so many organizations is that those who embrace a DevOPs methodology or are moving towards a DevOPs methodology are in a better place to manage this complexity.  The infographic below provides some things to consider…
DevOps_infographic http://www.dynatrace.com/en/initiatives/devops.html

Update December 1st, 2014 at 4:20PM

While Forever21 is not on any of our Retail Benchmarks,  http://www.isitdownrightnow.com/ caught Forever21 having an issue today. Screen Shot 2014-12-01 at 4.19.05 PM
Customers are understandably upset.
Screen Shot 2014-12-01 at 4.20.34 PM

Update December 1st, 2014 at 2:30PM

Andi Grabner has been looking at some details behind Staples slowdown using the Dynatrace Browser Agent. dtbrowseragent
It appears that a third party may be contributing to the slow down.  The Dynatrace Browser Agent is part of the Dynatrace Free Trial, and allows you to see from within the client where slow downs may be occurring. Give it a try. Dynatrace Free Trial

Update December 1st, 2014 at 2:15PM

CyberMonday traffic looks to be as high if not higher than Black Friday according to the Dynatrace Retail Index. index2pmcybermonday

Update December 1st, 2014 at 12:00PM

CyberMonday traffic is starting to claim some retail victims.  Staples has started to show a slow down. Screen Shot 2014-12-01 at 12.08.54 PM
Starting this morning after 8:00AM EST Staples site started to slow down.  The effects of this were seen from the Home Page, Mobile and Transactional tests that Dynatrace has been running.  The issues appear to resolve back to the Staples.com domain.  The Dynatrace Root Cause Analysis engine highlighted several areas of concern, but the primary area is long First Byte Times (this is time it takes waiting for a server to a request), which tend to point to issues in the Application. Screen Shot 2014-12-01 at 12.13.35 PM
While our testing uncovered this issue early this morning, we are also seeing reports filter in from Twitter users.
Screen Shot 2014-12-01 at 1.32.52 PM

Update December 1st, 2014 at 10:00AM

Traditional shopping patterns of the past are gone.  Shopping patterns have changed. Best Buy announced that they had to take their site down during Black Friday (http://www.cnbc.com/id/102223815) to reconfigure due to a spike in Mobile traffic.  Our Holiday Shopping survey told us to expect a change in the way people would be shopping this year.
Dynatrace 2014 Mobile Holiday Shopping Infographic (JPEG)
Now we are hearing reports in the media saying that in-store shopping numbers are down this year.  Based upon what we had seen during our survey, this is not surprising.  Mobile is allowing consumers to shop in unprecedented ways, allowing consumers to make purchases from anywhere including while they are in a Bricks and Mortar store.  So when a Retailer’s consumer is in a store, in an aisle, looking at a product on a shelf, with their mobile device in hand…  how important is the performance of that Retailer’s mobile site or app?

Update December 1st, 2014 at 8:45AM

Morning Everyone, So here are some quick results from a Dynatrace Retail Industry Index.   This is a sample of real user traffic coming from 35 different online retailers. weekend
Over the weekend, we saw that shopping interested start Wednesday evening, increased Thanksgiving morning and shot up on Thanksgiving after 4pm EST.  Black Friday morning is when we saw the most traffic and everything cooled down Saturday.  Finally we saw an increase starting last night leading into Cyber Monday.

Update November 28th, 2014 at 5:10pm

End of Day, Black Friday Wrap Up

Here we are at the end of the day on Black Friday.  As we expected Mobile traffic is changing the game this year.  Here are the top performing Mobile Retail sites this Black Friday. Screen Shot 2014-11-28 at 4.52.30 PM
For the Last Mile (these are real end user machines in peoples homes and offices) here are the top performers today. Screen Shot 2014-11-28 at 4.42.23 PM
It’s been a busy couple of days, and we are not done yet.  The holiday shopping season got off to very early start this year on Wednesday  evening, earlier than we ever seen it before. Check back with us on Cyber Monday.

Update November 28th, 2014 at 4:10PM

Andi Grabner had a closer look at Footlocker earlier today using Dynatrace for Browser Diagnostics. footlocker
24 MB of content for a mobile page (The Release Calendar Page) is a bit of stretch.

Update November 28th, 2014 at 3:45PM

When we look at our Retail  Industry Index we see some of the traffic start to ease off from earlier peaks. index traffic 330pm
This would be a good time to talk about some of the issues we’ve seen today.  Most of the major issues we’ve seen today were coming from the Application Side of these Retail sites.  We didn’t see anything to indicate there were malicious attacks happening or widespread network failures.  What we saw were sites going down because something inside that site failed to work as expected. The goal of online Retailers is to generate as much revenue as possible (that can be from online orders or supporting their in store shopping experience).  The goal of APM (Application Performance Management) vendors is to make sure that organizations like online Retailers can meet their goals as affectively as possible.  The applications which drive the online Retailers site are very complex.  Managing this complexity is not something that is done by a single group in those retailers organization, it is done through the cooperation of multiple teams (Line of Business, Development, QA, Production, etc…). The best way for those teams to work together is to agree upon a common set of metrics and a common set of tools.
Screen Shot 2014-11-28 at 3.09.09 PM
Above is an example of a simple visualization that allows multiple parts of an organization to have a lingua franca that they can share their understanding of how all of the interconnected parts work together to deliver functionality to the end user. Behind every site we track there is a map of how transactions flow, the servers will be different from Retailer to Retailer but the complexity is there and it must be managed.

Update November 28th, 2014 at 3:00PM

Time to have a look at how the third parties are doing this Black Friday afternoon.  As of 3:00PM EST Outage Analyzer is seeing 8 Open Outages. 7 of which are impacting US End Users.
Screen Shot 2014-11-28 at 2.55.42 PM
It appears that ev.ib-ibi.com and configusa.veinteractive.com are having some issues this afternoon. Screen Shot 2014-11-28 at 2.56.05 PM
In this case it does’t look as if they are slowing any sites down, however those site which are using there services are likely having issues with missing functionality.

Update November 28th, 2014 at 2:10PM

Earlier today BestBuy was down for a while.  Sources have told us that BestBuy brought their site down because they were seeing a spike in mobile traffic.  That is interesting because our own Holiday Shopping Survey told us to expect an increase in mobile traffic as more and more shoppers said they would be using their mobile device to shop. Mobile Holiday Infographic
So this isn’t surprising to us.  Managing mobile sites and apps requires new insight and new best practices for all organizations.  We track all of our Mobile Benchmark participants closely.
Screen Shot 2014-11-28 at 1.50.45 PM
As you can see we have been collecting a tremendous amount of data on BestBuy.  Let’s have a quick look at some of this data to see where Retailers need to pay attention to monitoring their mobile sites and apps.
Screen Shot 2014-11-28 at 1.40.22 PM
One of the first things we do is monitor the major underlying mobile carriers.  We have found that not all carriers operate in the same way.  Above you can see that the average response time for BestBuy’s mobile site across ATT, Sprint and Verizon is different. Let’s have a look at what may be causing that. Screen Shot 2014-11-28 at 1.40.48 PM
Above is a view of BestBuy’s mobile site, by carrier but we are looking at the amount of “over the wire” content (bytes) that are being delivered.  Unless BestBuy is delivering different content across each of these carriers the carriers themselves must be shapping the traffic.  This is not unexpected as carriers can employ different compression strategies to more efficiently push data to end users.  The challenge then becomes, how do we know what content is affected by this? Screen Shot 2014-11-28 at 1.49.06 PM
Above is a view of BestBuy’s Mobile Home Page but we’ve filtered out all of the third parties to focus on only www.bestbuy.com and images.bestbuy.com.  This view is also broken out by IP address, this allows us to see which CDN (Content Delivery Network) point of presence is delivering the content.  Since the CDNs have different relationships with the mobile carriers than BestBuy how do they know when a carrier is having routing issue from a particular CDN’s point of presence? The point I’m trying make here is that these sites are outrageously complex.  Today BestBuy had to close their shop to deal with an influx of Mobile traffic.  Retailers need more than ever as much visibility as they can get as to where that mobile traffic is coming from if they are going to be able to manage effectively. If they don’t then they will see more of this… Screen Shot 2014-11-28 at 2.13.20 PM

Update November 28th, 2014 at 1:00PM

Black Friday Retail Shopping Mobile and Web Analysis…  Some Details

PLEASE NOTE: I’m having some issues formatting the tables.  I have included some images of the tables Black Friday has been a challenge for some Retailers. Lets look at the details of the top and bottom performers and see if we can make some conclusions as to why they are where they are.

Mobile

This data is coming from the Dynatrace Mobile Synthetic Monitoring Network.  These are proactive tests taken from mobile web browsers from various locations and wireless carriers from across the US. With mobile sites it is particularly important to deliver the right kind/amount of data to that device for optimal experience.  Responsive Web Design is one way of making sure you are delivering the right content for the right device.  When we compare the Mobile results to the Last Mile results we see that some of the vendors (like Sears and Costco) have got it right, while other have not.

Top 10

Here are the Top 10 Mobile Retail sites as of 12:00pm EST on Black Friday. Screen Shot 2014-11-28 at 1.18.37 PM

Test

Response Time

Page Weight

Number of Objects

Number of Connections

Number of Hosts

Sears_MBHP-MA-iPhone-Retail-US – Mobile Agent – Mobile

0.87

59,101

8

3

2

Costco_MBHP-MA-iPhone-Retail-US – Mobile Agent – Mobile

1.62

302,526

18

6

6

Office Depot_MBHP-MA-iPhone-Retail-US – Mobile

1.66

192,519

17

16

11

REI_MBHP-MA-iPhone-Retail-US – Mobile

2.07

480,915

34

20

14

Saks Fifth Avenue_MBHP-MA-iPhone-Retail-US – Mobile Agent – Mobile

2.14

464,012

17

7

5

Tiger Direct_MBHP-MA-iPhone-Retail-US – Mobile

2.46

475,556

25

15

9

Amway_MBHP-MA-iPhone-Retail-US – Mobile Agent – Mobile

2.65

414,283

42

7

5

Walgreen_MBHP-MA-iPhone-Retail-US – Mobile

2.66

625,377

40

12

6

Dell_MBHP-MA-iPhone-Retail-US – Mobile Agent – Mobile

2.91

612,488

34

14

10

Newegg_MBHP-MA-iPhone-Retail-US – Mobile Agent – Mobile

3.16

560,638

26

17

12

Bottom 10

Here are the Bottom 10 Mobile Retail sites as of 12:00pm EST on Black Friday. Screen Shot 2014-11-28 at 1.21.49 PM

Test

Response Time

Page Weight

Number of Objects

Number of Connections

Number of hosts

L******************

6.98

721,192

94

54

32

V******************

8.05

651,972

117

73

39

A******************

8.67

5,587,109

53

6

4

F******************

8.68

776,989

156

61

42

N******************

9.23

673,694

104

81

40

C******************

11.84

1,274,531

204

76

49

J******************

13.38

248,076

135

90

45

E******************

15.19

2,125,296

212

154

79

N******************

17.14

2,331,209

275

170

85

P******************

21.10

2,474,787

102

34

19

So at quick glance it is not hard to see why some Mobile Retailers perform better than others.  As you increase the amount of content you are delivering (Page Weight in bytes) and increase the complexity of the site (number of objects and connections) and rely on more third parties (number of hosts), you will see a corresponding increase in Response Time.

Last Mile

The same pattern we saw from the Mobile Retail Benchmark is seen from our Last Mile Retail Benchmark.  Sites which are less complex perform much better than those that are more complex.

Top 10

Here are the Top 10 Retail Web sites as of 12:00pm EST on Black Friday. Screen Shot 2014-11-28 at 1.22.16 PM

Test

Response Time

Page Weight

Number of Objects

Number of Connections

Number of Hosts

Costco_LMHP-FF-Retail-US – Last Mile

2.59

1,813,060

103

16

6

Apple_LMHP-FF-Retail-US – Last Mile

3.19

3,178,981

56

16

6

Amway_LMHP-FF-Retail-US – Last Mile

3.39

786,555

70

20

12

Williams-Sonoma_LMHP-FF-Retail-US – Last Mile

3.96

2,770,274

114

48

31

PCConnection_LMHP-FF-Retail-US – Last Mile

4.16

634,164

97

21

10

Barnes and Noble_LMHP-FF-Retail-US – Last Mile

4.28

1,180,529

97

58

34

Dell_LMHP-FF-Retail-US – Last Mile

4.60

790,659

66

33

17

Crate and Barrel_LMHP-FF-Retail-US – Last Mile

4.62

1,503,148

112

73

49

JCPenney_LMHP-FF-Retail-US – Last Mile

4.70

1,323,980

149

79

47

Tiger Direct_LMHP-FF-Retail-US – Last Mile

4.80

2,481,974

249

46

28

Bottom 10

Here are the Bottom 10 Retail Web sites as of 12:00pm EST on Black Friday. Screen Shot 2014-11-28 at 1.23.22 PM

Test

Response Time

Page Weight

Number of Objects

Number of Connections

Number of Hosts

H******************

7.97

2,504,492

165

72

35

E******************

8.30

6,222,232

178

75

54

W******************

8.56

3,043,150

208

108

59

N******************

8.57

866,339

124

66

41

1******************

8.62

1,495,832

271

128

88

W******************

8.95

2,931,163

160

43

25

J******************

10.55

1,884,641

200

106

61

A******************

10.56

3,083,807

177

83

37

S******************

10.94

5,393,068

288

153

96

N******************

12.07

2,546,463

308

103

56

P******************

25.05

2,749,474

118

46

23

This afternoon we will look at a few of these sites in more detail to find out where they could be doing things better.

Update November 28th, 2014 at 12:02PM

I’m working on updating what we are seeing from the Retail Benchmarks, but in the meantime here is something interesting.  HP Shopping went down.  Given that HP was a major player in APM at one time it is really interesting to see this kind of performance. HP 2014-11-28 at 12.01.37 PM Here is what we were seeing from their home page. HP Screen Shot 2014-11-28 at 11.53.18 AM

Updated November 28th, 2014 at 11:20AM

BestBuy and Cabelas are back up.  That was fast turn around in getting their sites back up.

Updated November 28th, 2014 at 11:00AM

Electronics Retailer BestBuy is having some issues across the board. Let’s hope they can get them sorted out fast. BestBuy 2014-11-28 at 10.57.48 AM From a performance perspective they had been rock solid up until now with very consistent numbers. Bestbuy performance 2014-11-28 at 11.03.45 AM Since people are able to reach BestBuy.com it looks as though it is an internal error.  If it was an external issue, customers would most likely not even being seeing the “site is down” splash screen.

Updated November 28th, 2014 at 10:45AM

Last night we saw that Cabelas was having some issues. Today things look like they have gone from bad to worse for their online presence.  Here is a view of their page a few minutes ago. Cabelas Screen 2014-11-28 at 10.26.58 AM So when we look at the tests, we can see the extent of the problems.  Here is one of our Dynatrace Synthetic Performance Dashboards tracking Cabelas.  In this case the map is showing the result from the Dynatrace Last Mile network, these are real end user desktop machines in people homes. Cabelas Dashboard 2014-11-28 at 10.31.36 AM Definitely not a good situation to be in on Black Friday.  To be fair, Cabelas’s site is complex, there are a lot of moving parts. Cabelas TXN 2014-11-28 at 10.33.01 AM The Dynatrace Synthetic Performance Dashboard (above) is showing some of the complexity of that Cabelas has to manage.  For example the map in this view is showing where the content is coming from (from an end user point of view).  Cabelas uses a lot of Third Party and CDN services and that map shows exactly how complex their site is to manage. In the end when issues occur, regardless as to whether those issue are internal or external it is ultimately the end user who sees the results. cabelas twitter 2014-11-28 at 10.28.31 AM

Updated November 28th, 2014 at 10:20AM

Looks like FootLocker has been having some problems.  We are not sure if this was an intentional issue or an accident but when we look at the content being delivered (bytes) we can see when a change happened. footlocker 2014-11-28 at 10.10.06 AM Typically when investigating these issues we will look at where the content is being delivered from.  One of the unique capabilities of the Dynatrace Synthetic Monitoring network is it ability to look at a site’s hosts (domains) and filter them by where the content is being delivered from (actual IP address of where the content is being delivered from). footlocker detail 2014-11-28 at 10.08.22 AM Here we can see that FootLocker is deploying a CDN (Content Delivery Network).  Based on our experience the issue is likely a configuration issue between FootLocker and their CDN.  Again this is a very technical view of the problem, but these sites are very complex and difficult to manage. The end results speak for themselves.  Customers are having issues and tweeting about it. footlocker twitter 2014-11-28 at 10.11.05 AM What is really interesting is the real time site support that FootLocker is providing telling customers to try and refresh their pages.

Updated November 28th, 2014 at 9:45AM

Andi Grabner has been tweeting about some slow performance from JC Penny’s Mobile site. MobileWebAnalysis When we look at the overall results of the Benchmark participants we will most likely find that Retailers are being overly ambitous with what the are trying to server to end users.  The slow sites are going to be those that a content heavy and very complex.  Let’s see if these predictions play out. https://twitter.com/grabnerandi/status/538340542611193856

Updated November 28th, 2014 at 9:00AM

Black Friday Retail Shopping Mobile and Web Analysis

Once Again Dynatrace is closely watching Black Friday online retail site performance.  As we did last year we will focus on measurements from the Last Mile,(real end user machines) and Mobile .  These measurements are based on the last 24 hours.  We were anticipating to start by providing the last 12 hours average, however this year the shopping season started very early on Wednesday evening when we saw traffic increasing from our Retail Index and performance decreasing from our Retail Benchmarks. Our Holiday Survey discovered that end users would be depending on their mobile devices more this year.  Users were not just going to be couch shopping from their mobile devices at home, but bring them with them as they hit the stores.

Mobile

Lets start with the  top 10  fastest mobile  retail sites.   There has been some changes with who is on the list, but the standout is Sears, who remains a solid top performer from mobile devices this year and last year. This data is in average number of seconds it takes to load their home page from the Dynatrace Mobile Synthetic Network.  This network proactively runs tests from various locations and wireless carriers across the  US.

Top 10
Sears_MBHP-MA-iPhone-Retail-US – Mobile Agent – Mobile

0.85

Costco_MBHP-MA-iPhone-Retail-US – Mobile Agent – Mobile

1.61

Office Depot_MBHP-MA-iPhone-Retail-US – Mobile

1.66

REI_MBHP-MA-iPhone-Retail-US – Mobile

2.09

Saks Fifth Avenue_MBHP-MA-iPhone-Retail-US – Mobile Agent – Mobile

2.14

Tiger Direct_MBHP-MA-iPhone-Retail-US – Mobile

2.49

Amway_MBHP-MA-iPhone-Retail-US – Mobile Agent – Mobile

2.65

Walgreen_MBHP-MA-iPhone-Retail-US – Mobile

2.71

Dell_MBHP-MA-iPhone-Retail-US – Mobile Agent – Mobile

2.94

Newegg_MBHP-MA-iPhone-Retail-US – Mobile Agent – Mobile

2.99

Just to put the top 10 into perspective, we’ve include a list of the bottom performers as well.   We’ll be analyzing them further to find out what patterns  are making them perform so slowly.  Once we see those patterns we can look at simple strategies they can implement to improve their performance.

Bottom 10
L******************

6.98

V******************

8.21

F******************

8.41

A******************

8.99

N******************

9.11

C******************

11.77

J******************

13.59

E******************

15.01

N******************

17.12

P******************

21.17

We’ve hidden the names of these companies to add some mystery, but we will look at  some of them in detail later today.

Last Mile

With the  Last Mile, Dynatrace takes measurement from real end user machines.  These are machines running in people homes and offices.  We are seeing some of the same names in the Top 10 this year as we did last year.  For example  Apple  is once again in the  top 10.

Top 10

This data is the average number of seconds the home page takes to load from a random sampling of locations across the US.   Here are the top 10 fastest retails sites.

Costco_LMHP-FF-Retail-US – Last Mile

2.57

Amway_LMHP-FF-Retail-US – Last Mile

3.34

PCConnection_LMHP-FF-Retail-US – Last Mile

3.84

Apple_LMHP-FF-Retail-US – Last Mile

4.14

Barnes and Noble_LMHP-FF-Retail-US – Last Mile

4.21

Williams-Sonoma_LMHP-FF-Retail-US – Last Mile

4.22

Crate and Barrel_LMHP-FF-Retail-US – Last Mile

4.37

Dell_LMHP-FF-Retail-US – Last Mile

4.53

Macys_LMHP-FF-Retail-US – Last Mile

4.59

Tiger Direct_LMHP-FF-Retail-US – Last Mile

4.59

 Bottom 10

Companies can often overlook (sometime by mistake, sometimes by choice) factors which can cause their site to slow down.  To put the Top 10 in perspective here is what we are seeing from the bottom 10.

N******************

8.12

H******************

8.16

K******************

8.27

1******************

9.11

W******************

9.53

S******************

10.53

J******************

10.67

A******************

10.81

N******************

12.17

P******************

24.94

Again we are  masking the bottom performers, but will look into why they performing slowly later today. Check back throughout the day as we uncover what strategies the top performers are using and how they compare to the bottom performers.

Updated November 27th, 2014 at 11:30PM

Black Friday traffic is really starting to pick up.

Here we see a traffic index of 35 top retailers as 10pm EST. RetailPageViewsNov24-27midnight Twitter is heating up with people sharing their stories and experience while shopping online.  From an APM (Application Performance Management) point of view, by the time your customers are tweeting about an issue it is too late.  While APM tools can help you diagnose the issue, the damage is done.  Companies that are moving towards a DevOPs methodology stand a better chance of mitigating a performance issue if they have a set of tools that can be used across the organization. Whether it is your website or your mobile app, driving a performance culture allows you discover the issues we are seeing before they are getting tweeted about. Tomorrow we will be looking more closely at the Mobile performance of the Dynatrace Retail Benchmark.  Our Holiday Shopping study has shown that shopping behavior is changing as more and more people use their mobile devices to either do their online shopping or assist with their in-store shopping. Mobile Holiday Infographic Just to give you a taste for some of the data we are seeing here is Dynatrace Mobile Retail Benchmark for the past two days.  The Mobile sites look like they a faring well this evening, but we will have a closer look tomorrow. mobile Shot 2014-11-27 at 11.02.43 PM We will pick things up in the morning at 9:00AM EST with a Top10 of who is performing well on the Dynatrace Retail Benchmarks for Web and Mobile. We will also have a look at why the poor performers are faring so badly and what they could do to solve their  performance issues.

Updated November 27th, 2014 at 10:00PM

Changing The Rules

For most companies a “We’ll Be Back” message is a sign of something going wrong.  Apple is one of the few that can use it for dramatic effect. Apple 2014-11-27 at 10.05.35 PM

Updated November 27th, 2014 at 9:30PM

So Let’s Talk About Third Parties

Here at Dynatrace we not only pay attention to the sites being monitored but also with the Third Party components of those sites.  These can be CDN’s, ad networks, social media, analytic tools, etc…  Every third party provides some sort of functionality, but they also can introduce risk to a site in terms of performance and end user impact or client side errors.   Third Parties can cause a page to load slowly, or in the worst case, not at all.  Dynatrace collects millions of browser requests for these Third Party objects and aggregates them in our Outage Analyzer (http://www.outageanalyzer.com). When we look at Outage Analyzer this evening we see a number of Third Party events that are impacting multiple sites. Liverail 2014-11-27 at 9.22.59 PM The event that is impacting the most sites as of 9:30pm EST is coming from https://t4.liverail.com/ Screen Shot 2014-11-27 at 9.24.46 PM We can see that this single Third Party is potentially impacting close to 470 sites.  Six of those sites are on our Retail Benchmark (companies like REI and Victoria Secrets).  While this Third Party is not bringing those sites down, there are definitely objects errors which are being seen by end user browsers.  Even when we do a PING we can see issues with dropped packets. Liverail 2014-11-27 at 8.08.39 PM

Updated November 27th, 2014 at 8:00PM

So It Starts: Cabelas Having Some Issues

Cabelas has bee having some problems starting before 10am this morning. Cabelas Chart 2014-11-27 at 7.44.01 PMThe Dynatrace Synthetic Monitoring network immediately gave us where to start looking for issues with our Root Cause Analyzer (RCA).  Root Cause Analyzer looks through hundreds of tests and analyzes 1000’s of objects to look for patterns which would be difficult for an operator to find sifting through log files. Cabelas 2014-11-27 at 7.35.58 PMThe problem looks like a Socket Receive Timeout.  These are typically issues with network facing infrastructure like a load balancer or a web server.  While this may sound very technical it does have real world impact.  Our browser agent tests caught this issue and so did customers trying to get to Cabelas site (as we are seeing on Twitter). Cabelas Twitter 2014-11-27 at 7.34.31 PM

Updated November 27th, 2014 at 6:30PM

So looks like the holiday shopping season got off to an early start this year.  We’ve started to see activity as of last night at 6pm EST.  The Dynatrace Retail Home Page Benchmarks saw that sites started to slow just after 6pm EST with performance getting back to normal after 9pm.  These Benchmarks proactively test the top 50 Online retailers from a dozen cities across the US from the Dynatrace Synthetic Monitoring Network. RetailBenchmarksLoadTimeNov26-27
When we look at an index of over 30 leading online retailers we see that there was definitely was an increase in traffic (beyond what we see during normal shopping days) starting last night after 6pm. RetailPageViewsNov23-27
Keep coming back to see how things shape up this year.

Updated November 27th, 2014 at 4PM

We recently predicted that Black Friday 2014 will heavily rely on mobile, and end users will expect a seamless omni-channel experience from store, to Web site, to mobile site/ application. Additionally, unhappy users will vent their frustrations on social media, so e-retailers are heavily relying on the success of a fast, continuous omni-channel experience this weekend. Which retailers are going to embrace the opportunity, and which are going to disappoint end users and tarnish their reputation? We will be analyzing, tracking and reporting the best and worst sites of Black Friday using our free tools like Performance Tests, Outage Analyzer, and Benchmark reports. Check back here periodically for the latest web and mobile performance Black Friday news!

About The Author
David Jones
David Jones Director, Sales Engineering. APM Evangelist David Jones is the Director of Sales Engineering and APM Evangelism for Dynatrace. He has been with Dynatrace for 10 years, and has 20 years’ experience working with web and mobile technologies from the first commercial HTML editor to the latest web delivery platforms and architectures. He has worked with scores of Fortune 500 organizations providing them the most recent industry best practices for web and mobile application delivery. Prior to Dynatrace he has worked at Gomez (Waltham), S1 Corp (Atlanta), Broadvision (Bay Area), Interleaf/Texcel (Waltham), i4i (Toronto) and SoftQuad (Toronto).

Comments

  1. How are the average response times for the mobile sites calculated ? Is this just for the homepage or across all pages/user actions. ex: Product Details, Add to Cart etc.

    • David Jones David Jones says:

      Thanks for the question. The mobile tests are home page only. Our benchmarking practice does track transactional tests that touch the home page, product search, product detail, add to cart, checkout and login. We did not include much of that in our blog but that data is available from our benchmarks (www.dynatrace.com/benchmarks)

  2. how was dell.com analysed in the blog above? Can you specify which page were hit in the above test & other specifics like uptime etc.

Speak Your Mind

*

Do the math so we know you are human *