user experience

Twitter’s Crawl

In answering the question of Usability, "Can I use it?" the sub-category of Page Load plays an instrumental part. Twitter is one such excellent example of a web product with Poor Page Load Time.

image_thumb2A company can have the best product around, but if the pages are too sluggish, if the product suffers recurring outages, if the user-product interaction is varied and inconsistent, the product’s overall Usability can, and does, suffer.

Quick-UX provides for the rapid, simple and quantifiable assessment of a product’s User Experience (UX). Among the various components that define a product’s Usability, as well as Quick-UX‘s, are Accessibility, Consistency, Recognition, Navigation, and Page Load Time.

In answering the question of Usability, "Can I use it?" the sub-category of Page Load plays an instrumental role. Page Load, often obfuscated or connected with other perceived causes of a product’s dissatisfaction, ultimately, either positively or negatively, presents an unquestionable influence on a product’s overall Usability.

Example: Poor Load Time (value = 0.0)

Twitter is fast becoming, and for some already is, an essential communication tool.


Yet, Twitter earns a Page Load Time variable value of 0, due to its intermittent slow performance, but more so contributing to this value are the constant outages felt through the year, month after month.

If the page doesn’t load, if requested action takes an interminable amount of time, if the likelihood of the next user action failing is constantly looming, the overall Usability of a product takes a terrible toll.

In 2009, according to Pingdom, Twitter experienced a total of 20.82 hours of downtime.


Outages of Twitter were not isolated to merely the entire site being unavailable, but also consisted of sub-sections, or sub-features not working or resulting undesirable or unexpected behavior. Contributing to the pervasive problem of Page Load Time is both the inaccessibility of the product as well as the inability of the users to obtain key information (missing updates, etc) and other bugs leading to incomplete or otherwise incorrect Page Loads.

A Quick Study

I quickly examined and compiled a list of incidents that affected the Page Load Time of the Twitter product, distinguishing between total downtime, and partial downtime and information inaccessibility, based upon the public posts on Twitters blog.

I did my best to not double count any problems, but it was difficult since many of the problems occur so frequently, and it is often difficult to distinguish, from these status blog posts alone, between a persisting problem being experienced or fixed, from that of a new emergence of a similar or same problem. Furthermore, I also excluded the impact on Page Load Time arising from scheduled maintenance/downtime – periods of time over which the user expectation would be most aligned with the product’s promise of Page Load Time.

Some of my notes regarding my review of Twitter’s 2009 product Page Load Issues…


Dec 17

Site Outage

DNS records compromised

Dec 14

sms service unavailable

Dec 8

unplanned downtime

Dec 7

unplanned downtime

Dec 6

high rate of failwhales


Nov 30

Unplanned downtime

high error rate; tmp disabled list feature

Nov 23

elevated error rate

Nov 11

high number of errors

Nov 6

elevated errors


Oct 21

elevated error rate

Oct 18

network connectivity problems

Oct 13

account lockouts after username/pw change

Oct 12

errors and inability to tweet

Oct 7

Unplanned downtime


Sept 10

site slowness

Sept 9

secure connection failed issues


August 24

unexpected service interruption

August 16

Oauth and API problems

August 15

unexpected downtime

August 11

Site outage

August 6

Site is down

DOS attack

August 2

Search Down

problem coming from migrating data centers


July 10

site latency


July 5

restoring accidentially suspended accounts


June 15


problem w/ maintenance by provider


May 30

unscheduled downtime

fatal software error

May 28

unable to create new accounts

captcha problem

May 27

site latency

May 27

Unplanned downtime

May 22

search down

May 21

robot errors

May 20

user search unavailable

May 14

unplanned downtime

May 8

latency issues

resulting from a scheduled site maintenance


Apr 28

elevated error rate

fail whales

Apr 13

slow load times and high error rates

Apr 9

high latency

also fb not updating

Apr 7

high site errors; downtime/load issues

Apr 6

maintenance (no advance warning); downtime

Apr 6

errors; downtime

fail whales, robot pages; missing tweets

Apr 3

errors; downtime

fail whales, robot pages


Mar 16

unplanned maintenance

widespread slowness

Mar 4

problems logging in

Mar 2

power failure

degraded performance


Feb 18

latency issues

very long load times

Feb 14


db problem

Feb 11

Site down

db problem


Jan 20

site slow

slow load times

Jan 16


notified user of potential for more downtime


Dec 17

timeline delays and missing tweets

Dec 10

problem posting tweets to FB

problem resulting from FB latency issues


Nov 5

missing mentions


Oct 28

no dmsg emails

Oct 15

timelines 0.5h behind

Oct 8

timeline delays



Sept 16

missing tweets


Sept 14

missing tweets for some

Sept 4

short delivery delays

Sept 2

some tweets & followings delayed

small subset?


August 12

timeline delays


July 28

missing followers for new users


June 29

viewing other people followers/following disabled


June 16

unable to find new users

June 12

search delay

new tweets not being picked up by search

June 3

delayed followings

resulting from spam attack


May 13

timeline delays

hardware failure

May 4

search running behind

search not processing real-time


Apr 22

data inconsistencies


[still being fixed on the 27th]

Apr 22

missing user images

Apr 14

delayed search results

Apr 10

missing updates

Apr 6

missing avatars and dmsgs

Apr 2

not finding self in people search



Mar 18

missing tweets

db inconsistency, etc.

Mar 16

Delays on following and dmsgs

Mar 12

missing updates & actions

Mar 11


data inconsistencies (msg, counts, other data)

Mar 9

inbound sms delay


Feb 6

inconsistent follower/following counts

Feb 6

txt msg posting delays

problem w/ provider

Feb 2

Missing updates

Feb 2

missing self

new users missing from search


Jan 30

follower/following counts wrong

due to replication lag

Jan 19

slow search

search fell behind realtime due to maintenance

Jan 8

Delivery delays

tweets slow to appear in the timeline

Jan 6

Delivery delays

tweets slow to appear in the timeline


That said, a clear picture of the Page Load Time experience felt by the Twitter product’s user base quickly emerged.

Approximately 14% of all days in the year experienced delays and disruptions, directly altering the Page Load Time of the product. And, another ~10% of the year’s days experience pages loading with missing information, resulting in a total number of days experiencing disruption at around 24% of the year or 86 days! (note: there may be some day overlap that is not taken into account in these numbers)


Note: Data for December is complete (only goes through December 21, 2009)

Should Do & A Clear Flight Path

When using Twitter, tweets, responses, searches can and sometimes do occur quickly and without incident. However, with such consistency of problematic service, fail whales, site latency, etc. Twitter earns no more than a value of 0 for Page Load Time; but with a clear path to improvement…

  • first, focus on the reliability of the Page Load, drastically reducing downtime,
  • then, focus on the missing data and other inconveniences, some of which are touched upon in my table of notes above.


Over the next several weeks I will be providing real-world examples of Page Load Time values…

Poor Load Time (value 0) [Twitter, Twine]
Delayed Load Time (value 0.5) [Conversation Pieces]
Prompt Load Time (value 1) [Facebook]

Subscribe now (click here) to make sure you don’t miss any part of this series exploring the Usability and Page Load Time of Quick-UX, the quick and easy method of generating quantifiable and comparable metrics representing the understanding of the overall User Experience of a product, as well as other insightful posts from The Product Guy.

Enjoy, Discuss & Tweet!

Jeremy Horn
The Product Guy

Add to Social Bookmarks: Stumbleupon Furl Reddit Google Add to Mixx!
%d bloggers like this: