Skip to content

appYuser · Understanding the indicator

appYdex: the UX index for web performance

appYdex converts your visitors’ load times into a clear score from 0 to 1. Identify the pages and usage contexts affected by slow loading, share priorities across your teams and track the effects of your optimisations.

Real visit measurements · Apdex standard · Documented methodology

01 · The starting point

Apdex turns a time measurement into a score.

A raw load time is difficult to interpret on its own. Apdex compares it with a target time, then groups the measurements into three categories.

For an observed time To and a target time Tc, the categories are as follows:

  • Satisfactory: To ≤ Tc.
    The measured time is within the target.
  • Tolerable: Tc < To ≤ 4 × Tc.
    The time exceeds the target, but is no more than four times its value.
  • Frustrating: To > 4 × Tc.
    The time exceeds the tolerance threshold.
Example distribution of page loads across the satisfactory, tolerable and frustrating categories.
Example with a target time of 2 seconds: satisfactory up to and including 2 seconds, tolerable above 2 and up to and including 8 seconds, and frustrating beyond that.

A statistical indicator between 0 and 1

The calculation assigns a weight of 1 to satisfactory measurements and 0.5 to tolerable measurements. It divides this weighted total by the total number of measurements.

Apdex = (S + 0.5 × T) / N
S: satisfactory measurements · T: tolerable measurements · N: total number of measurements.

Apdex is an open standard. Quadran uses it as the basis for appYdex, combining two complementary Apdex scores.

Apdex formula: satisfactory measurements plus half of the tolerable measurements, divided by the total number of satisfactory, tolerable and frustrating measurements.
The calculation uses the collected measurements, grouped within the same scope and time period.

02 · Two complementary dimensions

One timing does not tell the whole story.

Consider two pages that finish loading at the same time. The first displays content quickly. The second stays blank for longer before appearing. Their total load time is identical, but the wait before the first elements become visible is different.

Initial rendering matters.

It tells you how long visitors wait before seeing the first elements of the page.

One timing is not enough.

The start of rendering and the full page load time each describe part of the experience when considered on their own.

The two complement each other.

appYdex takes both dimensions into account to provide a more complete view of load times.

Apdex for the start of rendering

It reflects the delay before the first visible elements appear. A slow start can degrade the experience, even when the total load time seems reasonable.

Apdex for full page load

It reflects the time needed to finish loading. Fast initial rendering is not enough to describe a page that continues loading for a long time.

Each Apdex is calculated using its own target time. The appYdex model then combines these two scores using the method described in its technical reference.

03 · Read the result

Five levels to assess performance.

appYdex ranges from 0 to 1. The levels below help you interpret the score displayed to two decimal places.

  • Excellent0.94 to 1
  • Good0.85 to 0.93
  • Acceptable0.70 to 0.84
  • Poor0.50 to 0.69
  • Unacceptable0 to 0.49

Levels documented in the appYdex methodology, for scores displayed to two decimal places.

A score of 0.90 does not mean that 90% of visitors report being satisfied. appYdex estimates one dimension of the experience from load times. It complements user feedback without replacing it.

04 · The appYdex model

A documented methodology based on Apdex.

Developed by Quadran for appYuser, the model combines Apdex for the start of rendering with Apdex for full page load. The chart shows how the two scores contribute to the result.

The technical reference sets out the underlying principles, the model’s zones and how it is constructed. This gives your teams an explicit method for understanding the score and interpreting changes over time.

The two horizontal axes represent Apdex based on full page load and Apdex based on start of rendering. The vertical axis represents the resulting appYdex. Each axis ranges from 0 to 1. Apdex based on full page load Apdex based on start of rendering
The combination of the two Apdex scores determines appYdex and its interpretation level.
  • Excellent
  • Good
  • Acceptable
  • Poor
  • Unacceptable

05 · From scores to action

Identify priorities. Measure progress.

appYuser collects load times during real visits through Real User Monitoring (RUM). appYdex gives technical and business teams a common reference point to focus their analysis and track improvements.

Find the pages to improve

Identify the pages, devices and browsers with the lowest scores. Focus your analysis on the contexts that adversely affect visits.

Check the impact of optimisations

After a fix or a production release, compare the score and load times across comparable periods and traffic, using the same target times.

Prioritise according to your objectives

Cross-reference performance with journeys and conversions to prioritise the pages that matter for purchases, enquiries or registrations.

06 · Frequently asked questions

Interpret appYdex correctly.

What does appYdex measure?

appYdex describes the experience associated with load times by combining the start of rendering and the full page load. For an overall analysis of website quality, it is complemented by an assessment of usability, content and accessibility.

What is the difference between Apdex and appYdex?

Apdex is an open standard that converts measured times into a score. appYdex is the indicator developed by Quadran for appYuser: it combines two Apdex scores using Quadran’s own combination method, described in the technical reference.

Are conversions included in the calculation?

The calculation is based on load times. Conversions, abandonment and journeys are analysed alongside it to inform priorities. A correlation between performance and conversions helps formulate hypotheses, but is not sufficient to establish a cause-and-effect relationship.

Is navigation without a full page reload included?

The current appYdex calculation covers full page loads. Navigation without a full page reload is included in appYuser journeys and sessions, but not in this score.

Do a one-off test and RUM provide the same information?

A one-off test examines a chosen situation. RUM observes the conditions of measured visits, when the tag can run. These approaches complement each other: a test result does not automatically represent the diversity of your traffic.

How can I compare the score between two periods?

Keep the same target times and check that the periods and visitor populations are comparable. Changes in the share of mobile traffic, the pages viewed or connection conditions can affect the score.

Explore further

Measure, then put the results to use.

appYuser

Web performance and digital experience

Discover how to use actual visitor data to guide improvements to your site.

Explore appYuser

Identify your optimisation priorities.

Discover how appYuser combines appYdex, journeys and conversions to inform your priorities.

Request an appYuser demo