Measuring the Business Impact
of Web Performance
1
Buddy Brewer [@bbrewer]
2
5
Why Care About Performance?
Slow load times damage traffic and sales
5
Why Care About Performance?
Slow load times also damage reputation
5
Performance isn’t just one simple number
Web sites deliver good and bad experiences every day
6
How Should I Approach Performance?
Performance optimization in three steps
How fast am I? How fast should I be? How do I get there?
7
How fast should I be?
The industry has lagged here
4
What is Real User Measurement?
JavaScript-based Real User Performance
Analytics
8
Tracking Conversions
What’s a “Conversion”?
Orders
Shares, Likes, Comments
Page Views
Subscriptions
Signups
Dealer Locators
Game Item Purchase
9
Measuring the impact of speed on business
Example 1: Impact of speed on conversion rate
9
Measuring the impact of speed on business
Example 2: Impact of speed on bounce rate
10
This Means We Can Measure Patience
And if we can measure it, we can manage it
11
Example
Time range: 1 month
Median Load Time: 4.12
Visits: 25M
Conversion Rate: 2.7%
Average Order: $100
12
Can We Do Better?
Speed increases drive business improvements
Median Load Time: 4.12s
Total Conversion Rate: 2.7%
Conversion Rate @ 3.0s: 4.9%
12
Faster pages convert more frequently
4.12 sec
2.7%
3.0 sec
4.9%
13
What Are We Playing For?
Impact is Measurable and Significant
Total Conversion Rate: 2.7%
Best Case Conversion Rate: 4.9%
Conversion Gap: 2.2%
Visits: 25M
AOV: $100
14
What Are We Playing For?
Impact is Measurable and Significant
(4.9% - 2.7%) * 25M * $100 =
$55M
15
What Are We Playing For?
Impact is Measurable and Significant
1 second = $55M
19
BUT
12
Potential vs Realistic Goals
Impossible to achieve ideal speed at 100th percentile
Median Load Time: 4.12s
Total Conversion Rate: 2.7%
Conversion Rate @ 3.0s: 4.9%
13
Refining Performance Goals
We can make these more realistic and iterative
Target Load Time: 4 sec (vs 3 sec)
Percentile at 4 sec: 49
Target Percentile: 60th
Percentile Gap: 11%
14
What Are We Playing For?
Impact is Measurable and Significant
(4.9% - 2.7%) * (11% * 25M) * $100 = $6M
Improving from
4.12 sec @ 50th percentile
to
4.0 sec @ 60th percentile
=
$6M / month
100% data collection
Stored forever
Processed in real time
Correlated to things that matter
© 2014 SOASTA CONFIDENTIAL - All rights reserved. 26
8
Customer Spotlight - Staples
Performance improvements lead to business gains
mPulse data led to changes in JavaScript, CSS
Reduced JavaScript size by 48%
Refactored CSS, reduced size by 83%
Shaved 1 full second off of Home page median load time
Shaved 6 seconds off the 98th percentile
Conversion improved by 10%
28
Free Speed Assessment
www.soasta.com/performance

How to the Measure Business impact of Web Performance

  • 1.
    Measuring the BusinessImpact of Web Performance 1 Buddy Brewer [@bbrewer]
  • 2.
  • 3.
    5 Why Care AboutPerformance? Slow load times damage traffic and sales
  • 4.
    5 Why Care AboutPerformance? Slow load times also damage reputation
  • 5.
    5 Performance isn’t justone simple number Web sites deliver good and bad experiences every day
  • 6.
    6 How Should IApproach Performance? Performance optimization in three steps How fast am I? How fast should I be? How do I get there?
  • 7.
    7 How fast shouldI be? The industry has lagged here
  • 8.
    4 What is RealUser Measurement? JavaScript-based Real User Performance Analytics
  • 9.
    8 Tracking Conversions What’s a“Conversion”? Orders Shares, Likes, Comments Page Views Subscriptions Signups Dealer Locators Game Item Purchase
  • 10.
    9 Measuring the impactof speed on business Example 1: Impact of speed on conversion rate
  • 11.
    9 Measuring the impactof speed on business Example 2: Impact of speed on bounce rate
  • 12.
    10 This Means WeCan Measure Patience And if we can measure it, we can manage it
  • 13.
    11 Example Time range: 1month Median Load Time: 4.12 Visits: 25M Conversion Rate: 2.7% Average Order: $100
  • 14.
    12 Can We DoBetter? Speed increases drive business improvements Median Load Time: 4.12s Total Conversion Rate: 2.7% Conversion Rate @ 3.0s: 4.9%
  • 15.
    12 Faster pages convertmore frequently 4.12 sec 2.7% 3.0 sec 4.9%
  • 16.
    13 What Are WePlaying For? Impact is Measurable and Significant Total Conversion Rate: 2.7% Best Case Conversion Rate: 4.9% Conversion Gap: 2.2% Visits: 25M AOV: $100
  • 17.
    14 What Are WePlaying For? Impact is Measurable and Significant (4.9% - 2.7%) * 25M * $100 = $55M
  • 18.
    15 What Are WePlaying For? Impact is Measurable and Significant 1 second = $55M
  • 19.
  • 20.
    12 Potential vs RealisticGoals Impossible to achieve ideal speed at 100th percentile Median Load Time: 4.12s Total Conversion Rate: 2.7% Conversion Rate @ 3.0s: 4.9%
  • 21.
    13 Refining Performance Goals Wecan make these more realistic and iterative Target Load Time: 4 sec (vs 3 sec) Percentile at 4 sec: 49 Target Percentile: 60th Percentile Gap: 11%
  • 22.
    14 What Are WePlaying For? Impact is Measurable and Significant (4.9% - 2.7%) * (11% * 25M) * $100 = $6M
  • 23.
    Improving from 4.12 sec@ 50th percentile to 4.0 sec @ 60th percentile = $6M / month
  • 25.
    100% data collection Storedforever Processed in real time Correlated to things that matter
  • 26.
    © 2014 SOASTACONFIDENTIAL - All rights reserved. 26
  • 27.
    8 Customer Spotlight -Staples Performance improvements lead to business gains mPulse data led to changes in JavaScript, CSS Reduced JavaScript size by 48% Refactored CSS, reduced size by 83% Shaved 1 full second off of Home page median load time Shaved 6 seconds off the 98th percentile Conversion improved by 10%
  • 28.