• Philip Tellis
• @bluesmoon
• philip@bluesmoon.info
• SOASTA
• boomerang
UX Fest 2013 / 2013-10-01 RUM Distillation 101 1
RUM Distillation 101
Philip Tellis / philip@bluesmoon.info
UX Fest 2013 / 2013-10-01
UX Fest 2013 / 2013-10-01 RUM Distillation 101 2
Did you know that most Rum comes from the
Caribbean?
UX Fest 2013 / 2013-10-01 RUM Distillation 101 3
But the largest selling Dark Rum in the world is
Old Monk
UX Fest 2013 / 2013-10-01 RUM Distillation 101 4
Jean-Antoine Nollet – 1746
UX Fest 2013 / 2013-10-01 RUM Distillation 101 5
Today Measuring the experience of Real Users is not
quite as shocking.
UX Fest 2013 / 2013-10-01 RUM Distillation 101 6
Real Users
UX Fest 2013 / 2013-10-01 RUM Distillation 101 7
Real Users
• Get delighted
• Get frustrated
• Have credit cards
• ... that they are attached to
• Form an impression from experience
UX Fest 2013 / 2013-10-01 RUM Distillation 101 7
Real Users
• Get delighted
• Get frustrated
• Have credit cards
• ... that they are attached to
• Form an impression from experience
UX Fest 2013 / 2013-10-01 RUM Distillation 101 7
Real Users
• Get delighted
• Get frustrated
• Have credit cards
• ... that they are attached to
• Form an impression from experience
UX Fest 2013 / 2013-10-01 RUM Distillation 101 7
Real Users
• Get delighted
• Get frustrated
• Have credit cards
• ... that they are attached to
• Form an impression from experience
UX Fest 2013 / 2013-10-01 RUM Distillation 101 7
Real Users
• Get delighted
• Get frustrated
• Have credit cards
• ... that they are attached to
• Form an impression from experience
UX Fest 2013 / 2013-10-01 RUM Distillation 101 7
What can we measure?
UX Fest 2013 / 2013-10-01 RUM Distillation 101 8
Anything that the user does!
UX Fest 2013 / 2013-10-01 RUM Distillation 101 9
Anything that the user does!
(in the browser)
UX Fest 2013 / 2013-10-01 RUM Distillation 101 9
But we really care about performance here...
UX Fest 2013 / 2013-10-01 RUM Distillation 101 10
So let’s talk about How
UX Fest 2013 / 2013-10-01 RUM Distillation 101 11
Modern Browsers – no problem
UX Fest 2013 / 2013-10-01 RUM Distillation 101 12
Modern Browsers – no problem
UX Fest 2013 / 2013-10-01 RUM Distillation 101 13
Real Users don’t always use the latest browsers.
UX Fest 2013 / 2013-10-01 RUM Distillation 101 14
So there’s a lot of magic to support the other
31.38%
UX Fest 2013 / 2013-10-01 RUM Distillation 101 15
The other 31.38%...
1 Get a timestamp before they leave a page
2 Save it in a cookie
3 Compare with timestamp when the page loads
UX Fest 2013 / 2013-10-01 RUM Distillation 101 16
The other 31.38%...
1 Get a timestamp before they leave a page
2 Save it in a cookie
3 Compare with timestamp when the page loads
Then there’s some magic to handle edge-cases
UX Fest 2013 / 2013-10-01 RUM Distillation 101 16
...magic
• What if the user goes off-site between pages?
• What if there’s SSL involved?
• What about new tab?
• What about those strange cases where they leave a page
after they load a new one?
• What about those common cases where they open
multiple pages from a search engine but don’t actually look
at them?
UX Fest 2013 / 2013-10-01 RUM Distillation 101 17
But it doesn’t need to be...
We put all this magic into boomerang
http://lognormal.github.com/boomerang/doc/
UX Fest 2013 / 2013-10-01 RUM Distillation 101 18
boomerang
• BSD licensed JavaScript library
• Measure load time, bandwidth, latency
• DNS lookup time, NavTiming, Connection Type
• Page Parameters
• And anything else you can write a plugin for
UX Fest 2013 / 2013-10-01 RUM Distillation 101 19
And it beacons all of this back to a predefined URL
UX Fest 2013 / 2013-10-01 RUM Distillation 101 20
So what then?
UX Fest 2013 / 2013-10-01 RUM Distillation 101 21
Are you Ops, F2E or Product?
UX Fest 2013 / 2013-10-01 RUM Distillation 101 22
Spot the datacenter
UX Fest 2013 / 2013-10-01 RUM Distillation 101 23
It’s actually in Nederland
UX Fest 2013 / 2013-10-01 RUM Distillation 101 24
How does bandwidth change around the world?
UX Fest 2013 / 2013-10-01 RUM Distillation 101 25
Did you know that IE7 seems to be faster than IE8?
UX Fest 2013 / 2013-10-01 RUM Distillation 101 26
But that’s mostly IE9 users browsing in IE7
compatibility mode.
UX Fest 2013 / 2013-10-01 RUM Distillation 101 27
What does a website’s performance distribution look
like?
UX Fest 2013 / 2013-10-01 RUM Distillation 101 28
That’s my Blog (please visit)
UX Fest 2013 / 2013-10-01 RUM Distillation 101 29
Log-normal distribution
UX Fest 2013 / 2013-10-01 RUM Distillation 101 30
Bimodal distribution
UX Fest 2013 / 2013-10-01 RUM Distillation 101 31
Distributions
• Sparse distribution suggests that you don’t have enough
data points
• Log-normal distribution is typical
• Bi-modal distribution suggests two (or more) distributions
combined
UX Fest 2013 / 2013-10-01 RUM Distillation 101 32
Distributions are interesting, but how does it affect
engagement?
UX Fest 2013 / 2013-10-01 RUM Distillation 101 33
Did you know that user patience wrt performance is
a cultural thing?
UX Fest 2013 / 2013-10-01 RUM Distillation 101 34
It’s tied both to where a user is from, and also the
type of service.
UX Fest 2013 / 2013-10-01 RUM Distillation 101 35
different types of sites
UX Fest 2013 / 2013-10-01 RUM Distillation 101 36
We borrowed a biological term called LD50 and
applied it to Load Time v/s Bounce Rate
UX Fest 2013 / 2013-10-01 RUM Distillation 101 37
LD50
UX Fest 2013 / 2013-10-01 RUM Distillation 101 38
This helps you fix only the stuff that’s hurting you.
UX Fest 2013 / 2013-10-01 RUM Distillation 101 39
Did you know that users bounce if a page is too
fast?
UX Fest 2013 / 2013-10-01 RUM Distillation 101 40
But that’s mostly because error pages are the fastest.
UX Fest 2013 / 2013-10-01 RUM Distillation 101 41
Measure → Hypothesize → Test → Fix
← Repeat ←
UX Fest 2013 / 2013-10-01 RUM Distillation 101 42
What delighted users a few years ago is now an
expected baseline, the absence of which will frustrate
them.
UX Fest 2013 / 2013-10-01 RUM Distillation 101 43
Changing Expectations
UX Fest 2013 / 2013-10-01 RUM Distillation 101 44
Thank You!
Questions?
UX Fest 2013 / 2013-10-01 RUM Distillation 101 45
• Philip Tellis
• @bluesmoon
• philip@bluesmoon.info
• www.SOASTA.com
• boomerang
• LogNormal Blog
UX Fest 2013 / 2013-10-01 RUM Distillation 101 46
Image Credits
• Jean-Antoine Nollet
http://en.wikipedia.org/wiki/Jean-Antoine_Nollet
• New York Times Olympic Medal Winners
http://www.nytimes.com/interactive/2012/08/05/sports/olympics/the-
100-meter-dash-one-race-every-medalist-ever.html
UX Fest 2013 / 2013-10-01 RUM Distillation 101 47

RUM Distillation 101 -- Part I

  • 1.
    • Philip Tellis •@bluesmoon • philip@bluesmoon.info • SOASTA • boomerang UX Fest 2013 / 2013-10-01 RUM Distillation 101 1
  • 2.
    RUM Distillation 101 PhilipTellis / philip@bluesmoon.info UX Fest 2013 / 2013-10-01 UX Fest 2013 / 2013-10-01 RUM Distillation 101 2
  • 3.
    Did you knowthat most Rum comes from the Caribbean? UX Fest 2013 / 2013-10-01 RUM Distillation 101 3
  • 4.
    But the largestselling Dark Rum in the world is Old Monk UX Fest 2013 / 2013-10-01 RUM Distillation 101 4
  • 5.
    Jean-Antoine Nollet –1746 UX Fest 2013 / 2013-10-01 RUM Distillation 101 5
  • 6.
    Today Measuring theexperience of Real Users is not quite as shocking. UX Fest 2013 / 2013-10-01 RUM Distillation 101 6
  • 7.
    Real Users UX Fest2013 / 2013-10-01 RUM Distillation 101 7
  • 8.
    Real Users • Getdelighted • Get frustrated • Have credit cards • ... that they are attached to • Form an impression from experience UX Fest 2013 / 2013-10-01 RUM Distillation 101 7
  • 9.
    Real Users • Getdelighted • Get frustrated • Have credit cards • ... that they are attached to • Form an impression from experience UX Fest 2013 / 2013-10-01 RUM Distillation 101 7
  • 10.
    Real Users • Getdelighted • Get frustrated • Have credit cards • ... that they are attached to • Form an impression from experience UX Fest 2013 / 2013-10-01 RUM Distillation 101 7
  • 11.
    Real Users • Getdelighted • Get frustrated • Have credit cards • ... that they are attached to • Form an impression from experience UX Fest 2013 / 2013-10-01 RUM Distillation 101 7
  • 12.
    Real Users • Getdelighted • Get frustrated • Have credit cards • ... that they are attached to • Form an impression from experience UX Fest 2013 / 2013-10-01 RUM Distillation 101 7
  • 13.
    What can wemeasure? UX Fest 2013 / 2013-10-01 RUM Distillation 101 8
  • 14.
    Anything that theuser does! UX Fest 2013 / 2013-10-01 RUM Distillation 101 9
  • 15.
    Anything that theuser does! (in the browser) UX Fest 2013 / 2013-10-01 RUM Distillation 101 9
  • 16.
    But we reallycare about performance here... UX Fest 2013 / 2013-10-01 RUM Distillation 101 10
  • 17.
    So let’s talkabout How UX Fest 2013 / 2013-10-01 RUM Distillation 101 11
  • 18.
    Modern Browsers –no problem UX Fest 2013 / 2013-10-01 RUM Distillation 101 12
  • 19.
    Modern Browsers –no problem UX Fest 2013 / 2013-10-01 RUM Distillation 101 13
  • 20.
    Real Users don’talways use the latest browsers. UX Fest 2013 / 2013-10-01 RUM Distillation 101 14
  • 21.
    So there’s alot of magic to support the other 31.38% UX Fest 2013 / 2013-10-01 RUM Distillation 101 15
  • 22.
    The other 31.38%... 1Get a timestamp before they leave a page 2 Save it in a cookie 3 Compare with timestamp when the page loads UX Fest 2013 / 2013-10-01 RUM Distillation 101 16
  • 23.
    The other 31.38%... 1Get a timestamp before they leave a page 2 Save it in a cookie 3 Compare with timestamp when the page loads Then there’s some magic to handle edge-cases UX Fest 2013 / 2013-10-01 RUM Distillation 101 16
  • 24.
    ...magic • What ifthe user goes off-site between pages? • What if there’s SSL involved? • What about new tab? • What about those strange cases where they leave a page after they load a new one? • What about those common cases where they open multiple pages from a search engine but don’t actually look at them? UX Fest 2013 / 2013-10-01 RUM Distillation 101 17
  • 25.
    But it doesn’tneed to be... We put all this magic into boomerang http://lognormal.github.com/boomerang/doc/ UX Fest 2013 / 2013-10-01 RUM Distillation 101 18
  • 26.
    boomerang • BSD licensedJavaScript library • Measure load time, bandwidth, latency • DNS lookup time, NavTiming, Connection Type • Page Parameters • And anything else you can write a plugin for UX Fest 2013 / 2013-10-01 RUM Distillation 101 19
  • 27.
    And it beaconsall of this back to a predefined URL UX Fest 2013 / 2013-10-01 RUM Distillation 101 20
  • 28.
    So what then? UXFest 2013 / 2013-10-01 RUM Distillation 101 21
  • 29.
    Are you Ops,F2E or Product? UX Fest 2013 / 2013-10-01 RUM Distillation 101 22
  • 30.
    Spot the datacenter UXFest 2013 / 2013-10-01 RUM Distillation 101 23
  • 31.
    It’s actually inNederland UX Fest 2013 / 2013-10-01 RUM Distillation 101 24
  • 32.
    How does bandwidthchange around the world? UX Fest 2013 / 2013-10-01 RUM Distillation 101 25
  • 33.
    Did you knowthat IE7 seems to be faster than IE8? UX Fest 2013 / 2013-10-01 RUM Distillation 101 26
  • 34.
    But that’s mostlyIE9 users browsing in IE7 compatibility mode. UX Fest 2013 / 2013-10-01 RUM Distillation 101 27
  • 35.
    What does awebsite’s performance distribution look like? UX Fest 2013 / 2013-10-01 RUM Distillation 101 28
  • 36.
    That’s my Blog(please visit) UX Fest 2013 / 2013-10-01 RUM Distillation 101 29
  • 37.
    Log-normal distribution UX Fest2013 / 2013-10-01 RUM Distillation 101 30
  • 38.
    Bimodal distribution UX Fest2013 / 2013-10-01 RUM Distillation 101 31
  • 39.
    Distributions • Sparse distributionsuggests that you don’t have enough data points • Log-normal distribution is typical • Bi-modal distribution suggests two (or more) distributions combined UX Fest 2013 / 2013-10-01 RUM Distillation 101 32
  • 40.
    Distributions are interesting,but how does it affect engagement? UX Fest 2013 / 2013-10-01 RUM Distillation 101 33
  • 41.
    Did you knowthat user patience wrt performance is a cultural thing? UX Fest 2013 / 2013-10-01 RUM Distillation 101 34
  • 42.
    It’s tied bothto where a user is from, and also the type of service. UX Fest 2013 / 2013-10-01 RUM Distillation 101 35
  • 43.
    different types ofsites UX Fest 2013 / 2013-10-01 RUM Distillation 101 36
  • 44.
    We borrowed abiological term called LD50 and applied it to Load Time v/s Bounce Rate UX Fest 2013 / 2013-10-01 RUM Distillation 101 37
  • 45.
    LD50 UX Fest 2013/ 2013-10-01 RUM Distillation 101 38
  • 46.
    This helps youfix only the stuff that’s hurting you. UX Fest 2013 / 2013-10-01 RUM Distillation 101 39
  • 47.
    Did you knowthat users bounce if a page is too fast? UX Fest 2013 / 2013-10-01 RUM Distillation 101 40
  • 48.
    But that’s mostlybecause error pages are the fastest. UX Fest 2013 / 2013-10-01 RUM Distillation 101 41
  • 49.
    Measure → Hypothesize→ Test → Fix ← Repeat ← UX Fest 2013 / 2013-10-01 RUM Distillation 101 42
  • 50.
    What delighted usersa few years ago is now an expected baseline, the absence of which will frustrate them. UX Fest 2013 / 2013-10-01 RUM Distillation 101 43
  • 51.
    Changing Expectations UX Fest2013 / 2013-10-01 RUM Distillation 101 44
  • 52.
    Thank You! Questions? UX Fest2013 / 2013-10-01 RUM Distillation 101 45
  • 53.
    • Philip Tellis •@bluesmoon • philip@bluesmoon.info • www.SOASTA.com • boomerang • LogNormal Blog UX Fest 2013 / 2013-10-01 RUM Distillation 101 46
  • 54.
    Image Credits • Jean-AntoineNollet http://en.wikipedia.org/wiki/Jean-Antoine_Nollet • New York Times Olympic Medal Winners http://www.nytimes.com/interactive/2012/08/05/sports/olympics/the- 100-meter-dash-one-race-every-medalist-ever.html UX Fest 2013 / 2013-10-01 RUM Distillation 101 47