Who are the top RUM analytics providers?

igrigorik · June 27, 2013, 10:58pm

SELECT vendor, sites, ROUND(ratio*10000)/100 as percent FROM (
  SELECT     
    COUNT(DISTINCT pageid, 1000000) as sites, 
    RATIO_TO_REPORT(sites) OVER() AS ratio,
    CASE 
      WHEN REGEXP_MATCH(url, "http[s]*://[www.|ssl.]*google-analytics.com/ga.js") THEN "Google Analytics"
      WHEN REGEXP_MATCH(url, "http[s]*://rum-static.pingdom.net/prum") THEN "Pingdom"
      WHEN REGEXP_MATCH(url, "insight.torbit.com/.*.js") THEN "Torbit"
      WHEN url CONTAINS  "aksb.min.js" THEN "Akamai"
      WHEN url CONTAINS "mpulse.net/boomerang/" 
        OR url CONTAINS "lognormal.net/boomerang/" THEN "SOASTA"
      WHEN url CONTAINS "/gomez" 
        AND url CONTAINS "rum" 
        AND url CONTAINS ".js" THEN "Gomez"
      WHEN url CONTAINS "boomerang.js" THEN "boomerang.js"
      WHEN url CONTAINS "/eum/rum.js" THEN "New Relic"
    END as vendor
  FROM [httparchive:runs.latest_requests] 
  GROUP BY vendor
  HAVING vendor IS NOT NULL
)
ORDER BY sites desc

Note: Akamai dynamically injects their RUM script at the time when the HTML is being generated (with customer defined sampling rate). As a result, the absolute count will vary (and won’t report all sites).

igrigorik · June 28, 2013, 2:23am

@souders all merged into a single query now! The percent column is the relative market share for all the sites that have RUM installed. Trying to figure out how to add an absolute % as well, but that’s turning out to be a bit more tricky… /cc @fhoffa

rviscomi · May 25, 2017, 4:25pm

Reran the query with 4/15/17 data:

Not much change with GA but Pingdom, SOASTA, and Akamai have grown substantially.

Would also be interesting to see the distribution of Alexa ranks for each vendor before and after…

Edit: Looking at the absolute numbers, GA now has fewer sites, but a similar % of vendorshare. Where did they go?

NicJ · May 25, 2017, 5:18pm

Was there a change in New Relic detection? Going from 6440 sites to 2 seems suspect.

paulcalvano · May 25, 2017, 5:47pm

Changing the NewRelic detection to:

WHEN url CONTAINS “js-agent.newrelic.com” THEN “New Relic”

appears to work -

andydavies · June 6, 2017, 2:20pm

To add a few more providers into the mix (as AppDynamics can be self-hosted it may under-report their numbers)

SELECT vendor, sites, ROUND(ratio*10000)/100 as percent FROM (
  SELECT     
    COUNT(DISTINCT pageid, 1000000) as sites, 
    RATIO_TO_REPORT(sites) OVER() AS ratio,
    CASE 
      WHEN REGEXP_MATCH(url, "http[s]*://[www.|ssl.]*google-analytics.com/ga.js") THEN "Google Analytics"
      WHEN REGEXP_MATCH(url, "http[s]*://rum-static.pingdom.net/prum") THEN "Pingdom"
      WHEN REGEXP_MATCH(url, "insight.torbit.com/.*.js") THEN "Torbit"
      WHEN url CONTAINS  "aksb.min.js" THEN "Akamai"
      WHEN url CONTAINS "mpulse.net/boomerang/" 
        OR url CONTAINS "lognormal.net/boomerang/" THEN "SOASTA"
      WHEN url CONTAINS "/gomez" 
        AND url CONTAINS "rum" 
        AND url CONTAINS ".js" THEN "Gomez"
      WHEN url CONTAINS "boomerang.js" THEN "boomerang.js"
      WHEN url CONTAINS "js-agent.newrelic.com" THEN "New Relic"
      WHEN url CONTAINS "spdcrv" THEN "SpeedCurve"
      WHEN url CONTAINS "nccrum.core.js" THEN "NCC Web Perf"
      WHEN url CONTAINS "cdn.appdynamics.com" THEN "AppDynamics"
    END as vendor
  FROM [httparchive:runs.latest_requests] 
  GROUP BY vendor
  HAVING vendor IS NOT NULL
)
ORDER BY sites desc

benhastings · May 16, 2018, 1:25pm

Looking for my own purposes and thought I’d share the results as of May.

I’m curious what others think about this:
For the sites I’ve inspected and companies I’ve worked with that explore RUM as a performance tool, many of them use Google Analytics AND something else - is this analysis a fair representation of what’s out there?

I’m going to experiment with the querying tool some more, but haven’t arrived at any different results with a simple reordering of the case statement.

benhastings · May 18, 2018, 5:19pm

Another conclusion/question -
I reran the query without the “HAVING vendor IS NOT NULL” and ended up with about 82% of sites having no observable RUM presence. That’s not terribly surprising, but it’s interesting when thinking about market penetration of paid-for performance monitoring solutions.

igrigorik · July 13, 2018, 11:19pm

Rerunning the query against latest tables (with larger origin count)…

SELECT vendor, sites, ROUND(ratio*10000)/100 as percent FROM (
  SELECT     
    COUNT(DISTINCT page, 1000000) as sites, 
    RATIO_TO_REPORT(sites) OVER() AS ratio,
    CASE 
      WHEN REGEXP_MATCH(url, "http[s]*://[www.|ssl.]*google-analytics.com/ga.js") THEN "Google Analytics"
      WHEN REGEXP_MATCH(url, "http[s]*://rum-static.pingdom.net/prum") THEN "Pingdom"
      WHEN REGEXP_MATCH(url, "insight.torbit.com/.*.js") THEN "Torbit"
      WHEN url CONTAINS  "aksb.min.js" THEN "Akamai"
      WHEN url CONTAINS "mpulse.net/boomerang/" 
        OR url CONTAINS "lognormal.net/boomerang/" THEN "SOASTA"
      WHEN url CONTAINS "/gomez" 
        AND url CONTAINS "rum" 
        AND url CONTAINS ".js" THEN "Gomez"
      WHEN url CONTAINS "boomerang.js" THEN "boomerang.js"
      WHEN url CONTAINS "js-agent.newrelic.com" THEN "New Relic"
      WHEN url CONTAINS "spdcrv" THEN "SpeedCurve"
      WHEN url CONTAINS "nccrum.core.js" THEN "NCC Web Perf"
      WHEN url CONTAINS "cdn.appdynamics.com" THEN "AppDynamics"
    END as vendor
  FROM [httparchive:response_bodies.2018_07_01_desktop]  
  GROUP BY vendor
  HAVING vendor IS NOT NULL
)
ORDER BY sites desc

rviscomi · July 26, 2018, 5:10pm

I would suggest one change from @igrigorik’s query above to use the requests dataset rather than response_bodies, which is 3x the size. Since the query doesn’t care about the responses themselves (just the URL) requests would do.

igrigorik · July 27, 2018, 7:28pm

Hmm, just realized that we’re missing some GA tags in above query (analytics.js, gtag.js)…

SELECT vendor, sites, ROUND(ratio*10000)/100 as percent FROM (
  SELECT     
    COUNT(DISTINCT page, 1000000) as sites, 
    RATIO_TO_REPORT(sites) OVER() AS ratio,
    CASE 
      WHEN REGEXP_MATCH(url, "http[s]*://[www.|ssl.]*google-analytics.com/ga.js") THEN "Google Analytics"
      WHEN REGEXP_MATCH(url, "http[s]*://[www.|ssl.]*google-analytics.com/analytics.js") THEN "Google Analytics"
      WHEN REGEXP_MATCH(url, "http[s]*://[www.|ssl.]*googletagmanager.com/gtag/js") THEN "Google Analytics"
      WHEN REGEXP_MATCH(url, "http[s]*://rum-static.pingdom.net/prum") THEN "Pingdom"
      WHEN REGEXP_MATCH(url, "insight.torbit.com/.*.js") THEN "Torbit"
      WHEN url CONTAINS  "aksb.min.js" THEN "Akamai"
      WHEN url CONTAINS "mpulse.net/boomerang/" 
        OR url CONTAINS "lognormal.net/boomerang/" THEN "SOASTA"
      WHEN url CONTAINS "/gomez" 
        AND url CONTAINS "rum" 
        AND url CONTAINS ".js" THEN "Gomez"
      WHEN url CONTAINS "boomerang.js" THEN "boomerang.js"
      WHEN url CONTAINS "js-agent.newrelic.com" THEN "New Relic"
      WHEN url CONTAINS "spdcrv" THEN "SpeedCurve"
      WHEN url CONTAINS "nccrum.core.js" THEN "NCC Web Perf"
      WHEN url CONTAINS "cdn.appdynamics.com" THEN "AppDynamics"
    END as vendor
  FROM [httparchive:requests.2018_07_01_desktop]  
  GROUP BY vendor
)
ORDER BY sites desc

Using the above, a big bump for GA…

benhastings · August 5, 2018, 6:38pm

I can’t see the inside view, but is this really accurately representing RUM by including all of the GA content? I know developers can create and send performance metrics through GA, but it seems like that’s more of a boundary case.

GA isn’t on the short list at all when I think about RUM (performance) in any context. I’d be happy to know otherwise.

NicJ · January 27, 2020, 7:42pm

Updated query:

Renamed Akamai to Akamai (Legacy RUM)
Renamed SOASTA to Akamai mPulse
Added the lux.js URL for SpeedCurve
Added CloudFlare’s RUM

SELECT vendor, sites, ROUND(ratio*10000)/100 as percent FROM (
  SELECT     
    COUNT(DISTINCT page, 1000000) as sites, 
    RATIO_TO_REPORT(sites) OVER() AS ratio,
    CASE 
      WHEN REGEXP_MATCH(url, "http[s]*://[www.|ssl.]*google-analytics.com/ga.js") THEN "Google Analytics"
      WHEN REGEXP_MATCH(url, "http[s]*://[www.|ssl.]*google-analytics.com/analytics.js") THEN "Google Analytics"
      WHEN REGEXP_MATCH(url, "http[s]*://[www.|ssl.]*googletagmanager.com/gtag/js") THEN "Google Analytics"
      WHEN REGEXP_MATCH(url, "http[s]*://rum-static.pingdom.net/prum") THEN "Pingdom"
      WHEN REGEXP_MATCH(url, "insight.torbit.com/.*.js") THEN "Torbit"
      WHEN url CONTAINS "aksb.min.js" THEN "Akamai (Legacy RUM)"
      WHEN url CONTAINS "go-mpulse.net/boomerang/" 
        OR url CONTAINS "lognormal.net/boomerang/" THEN "Akamai mPulse"
      WHEN url CONTAINS "/gomez" 
        AND url CONTAINS "rum" 
        AND url CONTAINS ".js" THEN "Gomez"
      WHEN url CONTAINS "boomerang.js" THEN "boomerang.js"
      WHEN url CONTAINS "js-agent.newrelic.com" THEN "New Relic"
      WHEN (url CONTAINS "spdcrv" OR url CONTAINS 'speedcurve.com/js/lux.js') THEN "SpeedCurve"
      WHEN url CONTAINS "nccrum.core.js" THEN "NCC Web Perf"
      WHEN url CONTAINS "cdn.appdynamics.com" THEN "AppDynamics"
      WHEN url CONTAINS "static.cloudflareinsights.com/beacon.min.js" THEN "CloudFlare"
    END as vendor
  FROM [httparchive:requests.2020_01_01_desktop]  
  GROUP BY vendor
)
ORDER BY sites desc

(note you have to run in Legacy SQL mode)

2020-01-01 results:

Row	vendor	sites	percent
1	null	4280567	57.46
2	Google Analytics	2994638	40.2
3	New Relic	142304	1.91
4	Akamai mPulse	14058	0.19
5	Pingdom	5795	0.08
6	CloudFlare	3585	0.05
7	AppDynamics	3192	0.04
8	Akamai (Legacy RUM)	3047	0.04
9	SpeedCurve	1940	0.03
10	boomerang.js	761	0.01
11	NCC Web Perf	19	0.0

rviscomi · January 27, 2020, 8:20pm

HTTP Archive integrates Wappalyzer detections upstream in WebPageTest and they have an Analytics category. It’s not exhaustive for web perf tools but it’s great because the detections are open source and community-maintained. I’d definitely recommend filing a PR to get any major RUM tools added.

Here’s an example of querying the Wappalyzer data, which lives in the technologies dataset:

#standardSQL
SELECT
  app,
  COUNT(0) AS sites,
  COUNT(0) / total AS percent
FROM
  `httparchive.technologies.2020_01_01_desktop`,
  (SELECT COUNT(0) AS total FROM `httparchive.summary_pages.2020_01_01_desktop`)
WHERE
  category = 'Analytics'
GROUP BY
  app,
  total
ORDER BY
  sites DESC

This query processes 1.12 GB, which is much smaller than querying the requests tables.

app	sites	percent
Google Analytics	3087580	72.13%
Google Analytics Enhanced eCommerce	185820	4.34%
Hotjar	177381	4.14%
Yandex.Metrika	134906	3.15%
Matomo	117786	2.75%
New Relic	116995	2.73%
comScore	84665	1.98%
TrackJs	79999	1.87%
Liveinternet	77979	1.82%
Quantcast	48154	1.12%
Crazy Egg	42600	1.00%
SiteCatalyst	31086	0.73%
Statcounter	30389	0.71%
Mixpanel	27669	0.65%
Segment	27241	0.64%
Visual Website Optimizer	20877	0.49%
Intercom	18662	0.44%
Optimizely	16710	0.39%
Gemius	16113	0.38%
Clicky	15133	0.35%
WP-Statistics	12453	0.29%
Chartbeat	10578	0.25%
Mouse Flow	10551	0.25%
Inspectlet	6315	0.15%
BugSnag	6299	0.15%

NicJ · January 27, 2020, 8:22pm

Also modified the boomerang.js query a bit to match a bunch of URLs we found from open-source users like /boomerang-n.n.n.min.js and /boomerang:

SELECT vendor, sites, ROUND(ratio*10000)/100 as percent FROM (
  SELECT     
    COUNT(DISTINCT page, 1000000) as sites, 
    RATIO_TO_REPORT(sites) OVER() AS ratio,
    CASE 
      WHEN REGEXP_MATCH(url, "http[s]*://[www.|ssl.]*google-analytics.com/ga.js") THEN "Google Analytics"
      WHEN REGEXP_MATCH(url, "http[s]*://[www.|ssl.]*google-analytics.com/analytics.js") THEN "Google Analytics"
      WHEN REGEXP_MATCH(url, "http[s]*://[www.|ssl.]*googletagmanager.com/gtag/js") THEN "Google Analytics"
      WHEN REGEXP_MATCH(url, "http[s]*://rum-static.pingdom.net/prum") THEN "Pingdom"
      WHEN REGEXP_MATCH(url, "insight.torbit.com/.*.js") THEN "Torbit"
      WHEN url CONTAINS "aksb.min.js" THEN "Akamai (Legacy RUM)"
      WHEN url CONTAINS "go-mpulse.net/boomerang/" 
        OR url CONTAINS "lognormal.net/boomerang/" THEN "Akamai mPulse"
      WHEN url CONTAINS "/gomez" 
        AND url CONTAINS "rum" 
        AND url CONTAINS ".js" THEN "Gomez"
      WHEN REGEXP_MATCH(url, r".*boomerang[\d\-\w\.]*(\.js)?$") THEN "boomerang.js"
      WHEN url CONTAINS "js-agent.newrelic.com" THEN "New Relic"
      WHEN (url CONTAINS "spdcrv" OR url CONTAINS 'speedcurve.com/js/lux.js') THEN "SpeedCurve"
      WHEN url CONTAINS "nccrum.core.js" THEN "NCC Web Perf"
      WHEN url CONTAINS "cdn.appdynamics.com" THEN "AppDynamics"
      WHEN url CONTAINS "static.cloudflareinsights.com/beacon.min.js" THEN "CloudFlare"
    END as vendor
  FROM [httparchive:requests.2020_01_01_desktop]  
  GROUP BY vendor
)
ORDER BY sites desc

Which bumps it up quite a bit:

Row	vendor	sites	percent
1	null	4280567	56.88
2	Google Analytics	2994638	39.79
3	New Relic	142304	1.89
4	boomerang.js	76162	1.01
5	Akamai mPulse	14058	0.19
6	Pingdom	5795	0.08
7	CloudFlare	3585	0.05
8	AppDynamics	3192	0.04
9	Akamai (Legacy RUM)	3047	0.04
10	SpeedCurve	1940	0.03
11	NCC Web Perf	19	0.0

NicJ · January 27, 2020, 8:43pm

Thanks Rick! I’ll take a look at updating Wappalyzer’s data too

rockeynebhwani · July 12, 2020, 11:57pm

Is it possible to add ‘Dynatrace’ RUM solution to this list? I believe detection can be based on “ruxitagent*.js”

paulcalvano · July 14, 2020, 5:05am

Updated the query to use Standard SQL instead of Legacy SQL. Also added DynaTrace’s RUM.

SELECT 
  sites,
  ROUND(sites / (SELECT COUNT(DISTINCT page) FROM `httparchive.requests.2020_06_01_desktop`) * 100 , 2) AS pct_sites,
  vendor
FROM (
  SELECT     
    COUNT(DISTINCT page) as sites, 

    CASE 
      WHEN REGEXP_CONTAINS(url, r"http[s]*://[www.|ssl.]*google-analytics.com/ga.js") THEN "Google Analytics"
      WHEN REGEXP_CONTAINS(url, r"http[s]*://[www.|ssl.]*google-analytics.com/analytics.js") THEN "Google Analytics"
      WHEN REGEXP_CONTAINS(url, r"http[s]*://[www.|ssl.]*googletagmanager.com/gtag/js") THEN "Google Analytics"
      WHEN REGEXP_CONTAINS(url, r"http[s]*://rum-static.pingdom.net/prum") THEN "Pingdom"
      WHEN REGEXP_CONTAINS(url, r"insight.torbit.com/.*.js") THEN "Torbit"
      WHEN url LIKE "%aksb.min.js%" THEN "Akamai (Legacy RUM)"
      WHEN url LIKE "%go-mpulse.net/boomerang/%" 
        OR url LIKE "%lognormal.net/boomerang/%" THEN "Akamai mPulse"
      WHEN url LIKE "%/gomez%" 
        AND url LIKE "%rum%" 
        AND url LIKE "%.js%" THEN "Gomez"
      WHEN REGEXP_CONTAINS(url, r".*boomerang[\d\-\w\.]*(\.js)?$") THEN "boomerang.js"
      WHEN url LIKE "%js-agent.newrelic.com%" THEN "New Relic"
      WHEN url LIKE "%spdcrv%" OR url LIKE '%speedcurve.com/js/lux.js%'THEN "SpeedCurve"
      WHEN url LIKE "%nccrum.core.js%" THEN "NCC Web Perf"
      WHEN url LIKE "%cdn.appdynamics.com%" THEN "AppDynamics"
      WHEN url LIKE "%static.cloudflareinsights.com/beacon.min.js%" THEN "CloudFlare"
      WHEN url LIKE "%ruxitagent%.js%" THEN "Dynatrace" 
    END as vendor
  FROM `httparchive.requests.2020_06_01_desktop`
  GROUP BY vendor 
)
WHERE vendor IS NOT NULL
GROUP BY vendor, sites
ORDER BY sites DESC

Here’s the results for June 2020 Desktop -

rockeynebhwani · July 14, 2020, 6:25pm

Thanks @paulcalvano for adding Dynatrace. I don’t know how to identify DataDog OR Raygun’s RUM solutions but it will good to add those as well.

I remember @andydavies saying on a thread that he has a long list of RUM providers. @andydavies - are you able to list provider names here so that we can expand this analysis?

jrharalson · July 14, 2020, 10:05pm

Certainly would be interested in @andydavies listing/thread.

In the meantime I’ve added Montis, Site24*7, Uptrends, and Raygun. I’d like to add Catchpoint later.

SELECT 
  sites,
  ROUND(sites / (SELECT COUNT(DISTINCT page) FROM `httparchive.requests.2020_06_01_desktop`) * 100 , 2) AS pct_sites,
  vendor
FROM (
  SELECT     
    COUNT(DISTINCT page) as sites, 

    CASE 
      WHEN REGEXP_CONTAINS(url, r"http[s]*://[www.|ssl.]*google-analytics.com/ga.js") THEN "Google Analytics"
      WHEN REGEXP_CONTAINS(url, r"http[s]*://[www.|ssl.]*google-analytics.com/analytics.js") THEN "Google Analytics"
      WHEN REGEXP_CONTAINS(url, r"http[s]*://[www.|ssl.]*googletagmanager.com/gtag/js") THEN "Google Analytics"
      WHEN REGEXP_CONTAINS(url, r"http[s]*://rum-static.pingdom.net/prum") THEN "Pingdom"
      WHEN REGEXP_CONTAINS(url, r"insight.torbit.com/.*.js") THEN "Torbit"
      WHEN url LIKE "%aksb.min.js%" THEN "Akamai (Legacy RUM)"
      WHEN url LIKE "%go-mpulse.net/boomerang/%" 
        OR url LIKE "%lognormal.net/boomerang/%" THEN "Akamai mPulse"
      WHEN url LIKE "%/gomez%" 
        AND url LIKE "%rum%" 
        AND url LIKE "%.js%" THEN "Gomez"
      WHEN REGEXP_CONTAINS(url, r".*boomerang[\d\-\w\.]*(\.js)?$") THEN "boomerang.js"
      WHEN url LIKE "%js-agent.newrelic.com%" THEN "New Relic"
      WHEN url LIKE "%spdcrv%" OR url LIKE '%speedcurve.com/js/lux.js%'THEN "SpeedCurve"
      WHEN url LIKE "%nccrum.core.js%" THEN "NCC Web Perf"
      WHEN url LIKE "%cdn.appdynamics.com%" THEN "AppDynamics"
      WHEN url LIKE "%static.cloudflareinsights.com/beacon.min.js%" THEN "CloudFlare"
      WHEN url LIKE "%ruxitagent%.js%" THEN "Dynatrace" 
      WHEN url LIKE "%rum.monitis.com%.js%" THEN "Monitis" 
      WHEN url LIKE "%site24x7rum-min.js%" THEN "Site24x7"
      WHEN url like "%uptrends.com%rum.%.js%" THEN "Uptrends"
      WHEN url like "%cdn.raygun.io/raygun4js/raygun.%.js%" THEN "RayGun"
    END as vendor
  FROM `httparchive.requests.2020_06_01_desktop`
  GROUP BY vendor 
)
WHERE vendor IS NOT NULL
GROUP BY vendor, sites
ORDER BY sites DESC

Here are the results for June 2020 Desktop

Row	sites	pct_sites	vendor
1	3393853	68.03	Google Analytics
2	147983	2.97	New Relic
3	127671	2.56	boomerang.js
4	15396	0.31	Akamai mPulse
5	7773	0.16	CloudFlare
6	5667	0.11	Dynatrace
7	5018	0.1	Pingdom
8	2949	0.06	AppDynamics
9	2514	0.05	Akamai (Legacy RUM)
10	1150	0.02	SpeedCurve
11	963	0.02	RayGun
12	886	0.02	Site24x7
13	166	0.0	Monitis
14	12	0.0	NCC Web Perf
15	1	0.0	Uptrends

Topic		Replies	Views
Which of the well-known Web Analytics providers are the most popular? Analysis	2	2573	October 30, 2013
Which sites are using multiple Analytics providers? Analysis	1	1677	July 9, 2013
Adoption of Bootstrap design Analysis	6	2252	May 15, 2020
What is the distribution of Domains per Page (Desktop vs Mobile) Analysis	5	2750	October 31, 2013
What is the addressable market of active web development? Analysis	1	1884	January 7, 2022

Who are the top RUM analytics providers?

Related topics