What a difference, Real time Mobile Performance testing vs. something else
by Peter
For today’s test I picked the following web site: http://mpulp.mobi/ It’s actually designed for Mobile so it should produce some good results. And it didn’t disappoint!
First here’s the results using our 3Pmobile.com performance web service. Several really important things to note here…
- Underneath the URL (first red arrow) you can see some icons. These are as follows:
- Cell strength (99 asu) Wi-Fi was disabled, GPS location was found and the cache was empty (that one’s really important otherwise the performance results are skewed)
- Secondly (the other red arrow) you can see the exact type of phone I used, the Carrier (Sprint) and even the OS rev (we’ve learned that it does make a difference when it comes to performance)
To compare our results I went to Blaze.io’s web site and plugged in the same URL and picked an Android phone to test it. Sadly I have no idea what carrier, what the Wi-Fi signal strength is, the OS rev or even where the device is located. However plowing on with the test reveals these results.
It’s really tough to see but the numbers are as follows:
- 25 requests
- 404.6kb
- 12.05 seconds
So let’s compare the two. Sprint downloads 495k in 9.4 seconds. Blaze’s Android downloads 404k in 12.05 seconds (probably had something in the cache, but then it should have got the page loaded even faster. Interesting.)
Summary…
On my Sprint phone (real device) in Denver Colorado http://mpulp.mobi loads 22% faster than the same test on the same “type” of phone. So whose right? Well even though I’m biased I’m going with the real device on the real carrier network with a known position and a cleared cache. If Blaze can show me why they’re results are more accurate then I’m happy to listen. Until then – they’ve got some catching up to do – literally (it’s faster on the actual device!)
Posted in: #mobile, #webperf, #wpo, Performance