iPhone vs. Android – 45,000 Tests Prove Whose Embedded Browser is Faster

iPhone 52% slower than Android, Android speeds past iPhone on 84% of sites*

UPDATE: Due to differences between the iPhone’s embedded browser vs the iPhones native browser the results may vary. Read More
UPDATE: Adjusted post title adding “Embedded Browser” to reflect recent findings.
UPDATE: Adjusted use of percentages

Key Findings:

  • iPhone 52% slower on average
  • Android was faster on 84% of the sites
  • Android even faster on non-mobile sites
  • iPhone 4.3 and Android 2.3 not much faster than previous rev

Conclusions:

  • Browser vendors optimized for benchmarks, not real sites
  • JavaScript Acceleration doesn’t impact most websites
  • Android’s edge expected to impact tablets even further

Browser performance is a big deal. Browser speed was a major bullet point – if not the top point – in practically every browser release this last year. In the mobile world, the latest iPhone version (4.3) and Android version (2.3) both focused on their improved JavaScript engine and faster browser. Browser performance is all the rage, and everybody says theirs is faster.


We set out to discover which mobile browser is truly faster – when used on real sites. Our goal was to measure the true mobile browsing experience, and see which device comes out ahead. 45,000 page loads later, this report summarizes our conclusions. We can now give a definitive answer to the question: which browser is really faster, from a user’s point of view?
The results surprised us.

First of all, we found that Android’s browser is faster. Not just a little faster, the iPhone was a whopping 52% slower. Android beat iPhone by loading 84% of the websites faster, meaning Safari won the race only 16% of the time. While we expected to see one of the browsers come out on top, we didn’t expect this gap.

Secondly, we saw that despite the optimized JavaScript engines in the latest iPhone & Android versions, browsing speed did not get better. Both Apple and Google tout great performance improvements, but those seem to be reserved to JavaScript benchmarks and high-complexity apps. If you expect pages to show up faster after an upgrade, you’ll be sorely disappointed. Read on to get more info about both findings, as well as additional comparisons such as WiFi vs. 3G and mobile sites vs. regular sites.

What Makes This Study Unique

Past comparisons usually focused on custom-created benchmarks, such as the SunSpider JavaScript benchmark. While useful, these benchmarks are very different than real world sites, and don’t reflect the actual user experience. This study measured the load time of 1000 real web sites, mimicking the experience users would get when browsing on their smartphones.

Other comparisons attempted to compare a small set of sites manually. However, we can’t draw conclusions from such small sets, and it’s hard to rely on the accuracy of manual measurement. Performance measurement is made up of many variables, and measuring the same page 5 times usually yields 5 different results. In this study, the large number of data points overcomes this variability. Having 9000 measurements on each device has the statistical strength to reliably say which browser is faster, across the different websites.

A Few Words About Methodology

The study was done primarily on iPhone 4 and Google Nexus S. The websites used were those of the Fortune 1000 companies. Each page was loaded multiple times and on different days, measured primarily over WiFi. For each device, we used the median load time for the comparison. The total number of tests was over 45,000.

This study was made possible through custom apps we developed, used to measure page load time on mobile devices. These apps run on the actual devices, load a page on demand, and measure how long it took. These agents are available as a free service to measure your own site on Mobitest.

Update: While Nexus S and iPhone 4 share similar CPU’s, measurements indicate iPhone 4 is running it at a 770Mhz speed. This was likely done to improve battery life at the expense of performance, and may account for some of the gap.

For the full details about our methodology, see the “Detailed Methodology” Appendix below.

iPhone 4.3 vs. Android 2.3


Our primary comparison was, as mentioned, the latest Android vs. the latest iPhone. We wanted to find out whose browser is faster, and got a resounding answer:

Android’s browser is faster. MUCH faster. On average, iPhone 4.3 was 52% slower than Android 2.3, with a median load time of 2.144 seconds vs. iPhone’s median load time of 3.254 seconds. Both median load times are generally fast, but keep in mind the test was done over a fast WiFi connection, and both the devices and network weren’t doing anything else.

Android was faster than iPhone in 84% of the tested websites, and iPhone beat Android in 16% of the races. This demonstrates Android wasn’t just faster overall, but rather provided a faster browsing experience 4 times out of 5.

Android’s edge completely disappeared when looking at mobile specific sites. These are sites that were modified to match the mobile user experience, and tend to be smaller and lighter. On mobile sites, Android was only 3% faster, with a median load time of 2.085 seconds vs. iPhone’s 2.024 – effectively the same. On non-mobile sites, iPhone was 59% slower, with an average load time of 2.180 seconds compared to 3.463 seconds.

Android’s dominance in handling non-mobile sites is especially important when considering tablets. Tablets use the same OS and similar hardware phones do. However, users expect the full experience on tablets, not the simplified mobile sites. This means Android’s edge will make an even greater impact.

JavaScript Speed Doesn’t Mean Browsing Speed

Comparing the last too OS versions of each platform resulted in one of the most surprising findings of this study. Both Android 2.3 and the recent iPhone 4.3 tout a dramatically improved JavaScript engine compared to their latest versions. Tests performed using the SunSpider JavaScript Benchmark proved that they indeed improved the JavaScript engine significantly, doubling its performance.

Therefore, we naturally assumed that the new versions will show significantly better load times. But we assumed wrong. When comparing iPhone 4.3 to iPhone 4.2 we saw no noticeable improvement, and Android 2.3 was only marginally faster than Android 2.2.

Comparing iPhone 4.3 and 4.2 yielded practically identical results. iPhone 4.3 was faster on 51% of the sites, but was 2% slower on average, with a median load of 3.253 seconds vs. 3.182 seconds on iPhone 4.2 The median gap between the two was exactly zero. We ran 9,000 measurements on each device, and the results were consistent – page loads remained the same.

For Android, since we couldn’t use exactly the same hardware, we compared the Samsung Galaxy S (running Android 2.2) to the Google Nexus S (running Android 2.3). Both devices are recent Samsung devices, and have very similar hardware. Android 2.2 averaged 2.370 seconds, and Android 2.3 averaged 2.144. While that is still a 10% boost, Android 2.3’s JavaScript engine is almost 40% faster than its predecessor. Android 2.2 still won 42% of the races, and the median gap between the two was merely 65 milliseconds. In other words, browsing speed didn’t change much.

To verify our hardware, we ran SunSpider on these devices, and got the expected (different) results:

  • iPhone 4.3: 3,978.9
  • iPhone 4.2: 10,303.9
  • Galaxy S (Android 2.2): 5,840.7
  • Nexus S (Android 2.3): 4,257.2

Our conclusion is that JavaScript performance doesn’t impact an average page load time. Apparently, JavaScript is already so optimized that it doesn’t play a big role in the time it takes to load a page. It’s likely that rich AJAX applications benefit from these improvements, but users should not expect their casual web surfing to move faster.

A secondary conclusion is that you optimize what you can measure. SunSpider created a predictable test, and Apple and Google did their best to get a better score. Measuring load time of real world sites is harder, in part due to lack of good measuring tools. As a result, the browsers scored higher on the expected test, but showed no progress in a surprise quiz.

WiFi vs. 3G

If performance is a big deal for browsers, it’s an even bigger deal for cellular providers. Reliability and speed seem to be all cellular providers talk about – besides having the latest phones. Most of our tests were performed over WiFi, to minimize the impact of the network on our results, but we wanted to get a feel for how 3G will impact those results.

To do so, we ran the iPhone 4.2 tests over 3G as well. The results were better than we anticipated. As expected, loading over WiFi was faster in 82% of the cases. However, the gap was surprisingly small – only half a second! The median load time over WiFi was 3.182 seconds, compared to 3.607 over 3G.

It’s worth emphasizing that we tested from a good reception area, and at night. We did so to eliminate noise, to maintain consistent results. The download speed over 3G was 5.95 Mbps (measured using speedtest.net), while in the middle of the day it’s usually about 1Mbps. This test therefore demonstrates the potential performance of browsing over 3G, but more tests should be run under different networks, hours and reception conditions.

Mobile Sites vs. non-mobile sites


One way to improve your website’s mobile browsing experience is to create a mobile site. This is a good practice for improving usability, and is usually lighter in resources – and thus faster to load. We were curious to see whether mobile websites indeed provide the expected performance boost.

Out of the 1000 test sites, 175 had a website customized for mobile. On average, mobile websites were loaded in 2.062 seconds, compared to 2.857 – a significant 39% gap. The difference was even greater on iPhone, where mobile sites were 60% faster (2.085 vs. 3.463). On Android, mobile sites were merely 7% faster (2.024 vs 2.180).

In general, smaller sites load much faster. We defined small sites as sites with 30 or less resources (images, scripts, etc.) and big sites as sites with more than 30 resources. On iPhone 4.2, small sites loaded in 2.193 seconds on average compared to 4.412 for big sites – more than twice as fast.

Summary

This study provided a lot of unexpected insights.
We assumed that similar hardware specs and the same WebKit foundation would make iPhone and Android’s browsers perform equally. We assumed that a faster JavaScript engine equals a faster browser. We assumed that 3G would be way slower than WiFi, even under good conditions.

All of these assumptions have been proven wrong when we actually measured those scenarios. Without measuring, you don’t know when and where you need to optimize. The SunSpider JavaScript benchmark pushed browsers to optimize JavaScript performance, because they could measure it. We hope that tools such as Mobitest would eventually result in a similar improvement to the browsing experience.

Appendix: Detailed Methodology

For test sites, we needed 1000 real websites. We chose to use the Fortune 1000 index as a good indicator for reasonable volume websites.

For clients, we used the latest devices to ensure a fair comparison. We used iPhone 4 for the iPhone measurements, and Google Nexus S for Android 2.3 measurements. For our Android 2.2 measurements we chose the Samsung Galaxy S, since it has almost identical hardware to the Nexus S, making the OS the primary difference. For network, we ran most of our tests over a high speed WiFi router, connected to the internet using a fast DSL connection. The fast internet helped reduce the variability caused by the network, and focus on the browser’s performance. The 3G tests were performed over the Bell Mobility HSPA network, and were made in good reception areas and at low usage hours (mostly at night).

To get the actual numbers, we measured each of the 1000 pages 3 times on each device. From each of the 3 runs, we saved the median scan. We repeated this test 3 times on different days, and filtered out results higher than 40 seconds or lower than 400 milliseconds, as those usually indicated network and server errors. We then chose the median result of the remaining scans on each page and each device. Using medians instead of averages helped us get a representative page load time, without being affected by network anomalies skewing the results.

To summarize, we ran 9000 tests over WiFi on each of these devices: iPhone 4.2, iPhone 4.3, Galaxy S (Android 2.2) and Nexus S (Android 2.3). We also ran 9000 tests on iPhone 4.2 over the Bell 3G network. Lastly, we ran 1000 tests on desktop browsers to separate mobile sites from non-mobile sites. The total number of tests performed for this study was 46,000.

The measurement itself was done using the custom apps, which use the platform’s embedded browser. This means WebView (based on Chrome) for Android, and UIWebView (based on Safari) for iPhone. Manual verification showed that page load performance of the embedded browsers, when properly configured, is effectively identical to the stand-alone browsers. The load times are calculated using the “Document Complete” callback from the browser, which is a standard way of measuring a web page’s load time. As mentioned above, the agents are now a part of a free service available at http://blaze.io/mobile/, and we encourage you to try it out.

To distinguish mobile sites from non-mobile sites, we loaded the same 1000 sites through IE8. We then compared the number of resources required to load the page on iPhone as compared to IE8. If the desktop browser required 30 additional resources or more, we flagged the website as mobile. Otherwise, it was flagged as “not mobile”, meaning it does not have a customized version for mobile.

Disclaimers

We know there’s no such thing as a perfect web page load measurement. The number of variables involved in loading a single page is astonishing, including DNS resolution, packet loss, web server load and many others. While we couldn’t control all the variables, we did try to minimize them where possible: We used high connection speeds and WiFi to reduce the network speed variable; used median load times to filter timeouts; and repeated the test on different days to address web server load. For the rest, we rely on the large number of samples to assure accuracy.

If you’d like access to discuss these results or our conclusions, or want access to the raw data for further analysis, please contact us at research@blaze.io, and we’ll be happy to share.

*This report is based on our own analysis leveraging the technology and methodology outline in this report. Blaze Software Inc. is in no way affiliated with Google or Apple.

Posted on March 17th, 2011
  • Pingback: Android browsing 52% faster than iPhone? Maybe, maybe not… - SlashGear

  • http://www.webpagetest.org/ Patrick Meenan

    Awesome testing and analysis (and surprising results). We’ve seen similar “impact” of javascript optimizations on the desktop side so it’s interesting to see that mobile is also largely not constrained by javascript performance. I’m sure you’ll be getting a lot of requests for the raw data from both sides (and independents) to figure out “why” Android is faster (is it simply the configuration of the number of parallel connections or something deeper in the stack?).

    • Guypo

      We did verify the devices were connected with the same speed to the router and web using the speedtest.net app on each device.
      If Android’s network stack interacts with it’s browser better and leads to faster page load times, I would still call that a faster browsing experience. In addition, if either device has better browser-hardware integration, I would suspect iPhone does, as they’re both provided by the same vendor.

  • Pingback: Google’s Android Browser Is Faster Than iPhone’s Safari » Geeky Gadgets

  • Pingback: Chrome pe Android 2.3 este mai rapid ca Safari din iOS 4.3

  • Pingback: Android browsing 52% faster than iPhone? Maybe, maybe not… | Digital Tablets

  • Pingback: Android Vs iPhone Browser Speed Test Reveals Android is Quicker - Gadget Venue

  • Pingback: Report: Browsing the web on Android is 52% faster than iPhone | MobileSyrup.com

  • http://twitter.com/sbrocket Bryan Henry

    Interesting. Might be relevant when 2.3 actually shows up on phones other than Google’s own, but 2.2 has just reached over 60% adoption after 9 months with another 29% on 2.1. I don’t see 2.3’s 1% chunk expanding anytime soon. The major device manufacturers seem to have decided that 2.2 is what they want to target, newer versions be damned.

    • http://www.facebook.com/wademealing Wade Mealing

      Its not a question of adoption, but the versions.

  • alan

    Interesting results. Yesterday there were stories out that Web apps saved to iPhone home screens do not benefit from several of the improvements in the latest Safari (Nitro javascript engine, some kind of rendering module, and a third I can’t recall). Have you checked if using UIWebView inherits the new features? Obviously there are multiple ways to use WebKit within iPhone.

  • Anonymous

    nice attempt to make your unscientific approach seem legitimate by babbling on and on

    this article could have been 75% shorter
    its clear you are loading up the BS due to lack of “real” evidence

    benchmarks beat some human with a stopwatch and an agenda

    this article is bullshit

    • http://twitter.com/parallelogram parallelogram

      As usual an emotional reply from someone who seems to be an ardent Apple fan
      Learn not to fall in love with a giant tech company like it looks after your well being

      • Anonymous

        webOS fanboy here, actually

        rocking a SR-71 custom kernel

      • Jess

        As usual, an uneducated reply from someone who just assumes they know everything, and judges people from their high horse. Awesome work, parallelogram!

  • Pingback: Android 52% faster, speeds past iPhone on 84% of sites | PDA-247

  • Pingback: Browser do Google no Android é mais rápido que o Safari no iPhone | Geeklândia

  • Pingback: iPhone Slower Than Android In Loading Web Pages | WebProNews

  • Ramesh18 R

    Good review. I used both and I, as a normal user don’t find much difference.

    • A358303

      Wow look at all the hate against each other. : P
      Honestly just make your own choice and just pick one.

  • Pingback: Navegador padrão do Android é bem mais rápido que do iPhone | Gizmodo Brasil

  • Kobuspost

    Did you also run the Wifi vs 3G tests on Android?

  • Pat

    The conclusions of this test are critically flawed. It has been well documented in recent days that iOS 4.3’s UIWebView does NOT support the nitro engine for javascript acceleration. It therefore does no surprise that the difference between load times in iOS 4.2 and 4.3 are exactly the same – because they are using exactly the same engine! It would be interesting if you could repeat the test using MobileSafari itself to get the true picture.

    • jbrandonf

      You ignored the part about javascript improvements not impacting browsing speeds in real world usage.

  • Pingback: 45 000 tests prove Android browser is faster | Droid Sector

  • http://twitter.com/philipmcdermott Philip McDermott

    “Therefore, we naturally assumed that the new versions [with faster Javascript engines] will show significantly better load times. But we assumed wrong.”

    This is not a natural assumption: it’s an assumption made if you know nothing of web design.

    Whilst it may be the case that the Android browser renders pages faster than Safari mobile, this experiment absolutely does not prove it to be the case. You have no idea whether or not the difference was because the Android devices had either a) a faster connection to your Wi-Fi router, or b) have a faster network stack implementation.

    • Anonymous

      Sure, thats fair. What is really being compared here is one phone’s browser versus another phone’s browser. Or, maybe, one phone’s browser in a custom app versus another phones browser in a custom app. Still… the phones are fairly comparable and this is a useful thing to do, isn’t it?

  • http://indtechie.com Sidharth Dassani

    I am just waiting for John Gruber and MG Siegler to do their own tests to prove that the above results are totally false.

    • Anonymous

      Gruber and Siegler? Really unbiased opinions, indeed… Why not a Steve Jobs test too?

  • Pingback: iPhone vs. Android – 45,000 Tests Prove Whose Browser is Faster (Dave Horne/Blaze.io) | Webhosting Hosting

  • Pingback: Androidiani: Android più veloce del 52% di iPhone nella navigazione Internet

  • Eldad Eilam

    Here’s why this study is wrong: Because you used your own app (using UIWebView), and not using the Safari application. iOS 4.3 has a known flaw where a different (SLOWER!!) JavaScript engine is used when a 3rd party app uses the UIWebView object, which you MUST be using since you’ve developed your own app. If you had tested on the actual browser, things would have gone much differently.

    I am sorry to say this, but your results are essentially useless for any meaningful comparison. I’d suggest that you wait for Apple to fix this flaw and then rerun the iOS tests. THEN we’ll learn what’s what.

    • http://www.androidnews.de/ Frank (androidnews.de)

      “Manual verification showed that page load performance of the embedded browsers, when properly configured, is effectively identical to the stand-alone browsers.”

    • jwb

      The fact that Apple has provided a surprising, weird, and broken API for UIWebView, while Android provides full web browser performance to applications, is not at all irrelevant. Tons of applications use an embedded web browser, including Facebook and Yelp to name just a few.

      • Anonymous

        OK, make the headline read, “Blaze Software’s home-made browser on Apple isn’t as good as its home-built browser on Android.”

        Truth is, there are many other differences between Safari and the WebUIKit — get the “Kit” part? — than just the special permissions that Apple, likely for security reasons, does not grant to 3rd-party apps.

        Get many page views that way?

        For that matter, what was Blaze’s legitimate interest in burning all those engineers’ hours? To show how incompetent they are about testing browsing? (They’ve since filed an addendum to say that they are “looking into” one of the obvious flaws in their test that others reported.)

        Dishonest? Incompetent? Both? Take your pick.

      • Smdish60

        READ THE TITLE! It says specifically ‘Tests prove whose BROWSER is faster’ not 3G speed or hardware speed… stop concealing yourself in a shell…

    • http://twitter.com/SurreyStore Randy Hans

      You can easily install android on iPhone with iAndroid, find out how
      here
      http://www.surreystore.com/cms/apple/69-use-andiod-os-on-your-iphone.html

  • Pingback: Android Browser 52% Faster Than The iPhone In Speed Tests | TechnoBuffalo

  • Guypo

    Re being biased, I assure you we’re not. I personally own an iPhone, iPad and Mac, and love the apple design. They were just slower.

    Re nitro, we did our best to enable it, but can’t say for sure we did. However, the android results indicate JavaScript only makes up ~15% of the page load time. This means nitro could at most narrow the gap.

    • Anonymous

      Let’s be clear. “They” means “our home-made software” was just slower. You did NOT invoke the modern JIT built into Nitro, and you did NOT come anywhere close to invoking Safari’s specific asynchronous requests, but instead used the WebKit building blocks in a straightforward, if naïve, way.

      Why did you publish this study in the first place? Burn all those engineer hours? It’s certainly going to get you a reputation for being either incompetent or sloppy, no matter how much you profess you were purely disinterested.

      • Vumanhle1965

        Omg, get over it dude…its faster…life goes on…

  • VeMeth

    It’s so funny how the apple fanoboys come forward with agressive comments about how flawed and bad this test was. Says a lot about the user base.

    • dan

      It says nothing about the user base. Fan boys are a vocal minority. Best thing to do is tune them out.

    • Anonymous

      Damn tooting. OTOH, you apparently think any claims that Android is 50% faster, no matter how cooked up, is perfectly OK?

      • Mattharader

        Lol you must work for apple or something cause you are blind

        • Anonymous

          No, I’m an “independent contractor.” But while the pay is good, you have to work pretty hard. I understand Google pays people to make the sort of juvenile “you must work for Apple” and the fake macho “real tech types would never stuff for iSheeple” but I actually have to include real facts that’ll stand up to scrutiny.

          Still, it’s fun to tweak people who only know what the Google, Verizon and Adobe marketing departments disseminate. As long as I don’t think about how it’s sad.

  • Andrew

    Out of curiosity, you say that you use a custom app with an in-app browser and manually verified that page loading times were identical between your custom app and the native browser. However, Ars Technica and other sources say that the JavaScript improvements in iOS 4.3 don’t carry over to in-app browsers.

    http://arstechnica.com/apple/news/2011/03/confirmed-some-web-apps-not-seeing-ios-43-javascript-speedup.ars

    Did you account for that?

    • jbrandonf

      They also made a point about how for all the JavaScript improvements that Apple and Google were toting that it generally doesn’t translate into real world speeds.

  • Anonymous

    I agree with other posters that this is too long winded of an article. But the real problem is that you are comparing apples to oranges since the phones have different hardware specs. Certainly both have a 1GHz processor, but actual clock speed may be different. And don’t forget that the retinal display on the iPhone is significantly higher than any of your Android models. There are also OS settings than could be different between the two that impacts results. The bottom line is that this is not scientific. A true comparison of the browser speeds would require exact same hardware and system configuration.

    • think

      you can’t be serious. they chose nexus s because it has more or less the same hardware.
      what would you expect? a test of android’s browser on the iphone??

      • Blazeio

        That is precisely what I expect from a proper study. If the Android platform manages to “win” even on the hardware of the iPhone 4, I would consider it to have truly beaten the iPhone, especially given the fact that it’s reasonable to assume Apple has spent more effort optimizing iOS than a custom ROM developer did in porting Gingerbread to the iPhone 4.

        Show me the load times on identical hardware, with the measured-to-be-faster platform not playing a home game, and I would consider it to be a much more reliable study.

        And for the record: I’m an Android fanboy, I loathe iPhones. But this approach is just unfair to Apple.

        • robzr

          The iPhone 4 is mid product cycle (9 months old), runs at 777 MHz, the Nexus S is early product cycle (2 months old) and runs at 1 GHz. Thats apples and oranges. If they really wanted to be fair, they’d at least have used a 1st gen iPad which run at the same clock speed (1 GHz) as the Nexus S. Both having A8 cores that would at least be comparable.

          This study is ridiculous, but I’m SURE that Blaze will redo it when the iPhone 5 running the dual core A9 based A5 comes out, and using the Nitro engine, takes a second off the Android times. Right? Or would that bias not play into Blaze’s pocketbook?

          Rob

      • Anonymous

        Go and check every single kind of comparison for Apple Iphone made against an Android phone on every test,be it bench marking,usability,hardware specs,browsing speeds and,alot more kind….and apple fan boys comments of dissatisfaction made it more obvious they have been left behind and they simply couldn’t swallow the truth…..the bitter truth of their expensive I phone is not worth it.

  • mualexander

    Interesting findings — it would be great to see similar ones for Xoom vs. iPad 2.

    One nit, however, is the use of percentages. If A=2 and B=3, you can say that B is 50% more than A, but you can’t say that A is 50% less than B.

  • Turver

    I think the headline is misleading. Doesn’t the test show that Android’s _browser_ is faster than iPhone’s browser? Yet, the headline says: “Android faster than iPhone.”

    • Shanghai Dan

      Uhhh… English isn’t a primary language for you?

      iPhone vs. Android – 45,000 Tests Prove Whose Browser is Faster

      Seems to be quite explicit it’s talking about which BROWSER is faster…

      • http://www.facebook.com/profile.php?id=594289803 Jerry Ballard

        “45,000 Tests Prove Whose Browser is Faster”

        That is PRECISELY what they did NOT test.

    • robzr

      Very misleading, when the clock speed of the processor on the 9 month old iPhone 4 is 777 MHz vs the 1 GHz clock speed on the 2 month old Nexus S. Why didn’t they use the iPad 1, also running at 1 GHz if they truly were trying to compare the OS?

      Rob

      • Ann

        The question is which phone is fastest today, not which phone is fastest seven months ago or five years ago. Your complaint is frankly laughable.

        • robzr

          The article is comparing OS’s, not phones or hardware. To compare the OS, the hardware needs to be identical, or as identical as feasible. A phone that runs a similar CPU yet at a 28% faster clock speed is far from identical.

          Funny, the clock speed is 28% faster and the page loads were 34% faster. Kind of closes the gap a bit, doesn’t it.

          Now if this was a article about which PHONE renders webpages faster, then I probably would never have read it, because theres nothing interesting about new phone to one approaching the end of it’s life cycle. I would hope the Nexus S is faster, otherwise that’d be pretty embarrasing.

          Rob

  • Anonymous

    It is also interesting that your tests only loaded each site 3 times and used median instead of average. You argue that this will help to eliminate anomalies such as a network connection issue. But that is a false premise to justify using the median. What you should have done is to load each page many more times, eliminate outliers and do yhe average of each page. It would be interesting to know breakdown of which of the 3 page loads that the median came from on average for each of the phones. That itself might also tell you something.

    • Mattharader

      45,0000 is not enough? Reporting the average could be misleading because it does not tell which speed is most consistent. for example is someone scored 10 points in one basketball game but only 1 point the next other two games the average is 4 but the consistency is only 1. I do not want my mobile device to sometimes be excellent I want it to be excellent consistently.

  • http://www.mwd.com @JoeHobot

    Some people drank too much Apple Juice :) , its faster get over with it, whether the charts above are true or false…open up the browser and see it for your self.
    God dang it. :)

  • http://twitter.com/mschmidlen michael e schmidlen

    What’s “interesting” about this little “study” is the COMPLETE lack of disclosure about WHO is behind the boys @ the company/product formerly known as BlazeIO.

    Things that make you go Hmmmm.

    • http://www.blaze.io David Horne

      Hi Michael,

      I would be happy to fill you in. Blaze is a new startup located in Ottawa, Canada our core product is an application that automates front end performance optimization. We developed http://www.blaze.io/mobile as a testing tool to further our expertise in optimizing mobile devices.

      Because we had the tool available to measure performance on mobile devices we thought it would cause a lot of interest to release our findings.

      We have no association with Google or Apple.

      • Anonymous

        Which associations do you have?

    • Tinthing

      Well just click Company and Team and there are your names.. Click contact and give them a call.

      You didn’t check the website did you?

  • User

    When testing on 3G was WAP Proxy turned off for Android phones?

  • http://twitter.com/lleaf David Goldstein

    “The measurement itself was done using the custom apps, which use the platform’s embedded browser. This means WebView (based on Chrome) for Android, and UIWebView (based on Safari) for iPhone. Manual verification showed that page load performance of the embedded browsers, when properly configured, is effectively identical to the stand-alone browsers.”

    Given the recent discussions that the UIWebView for IOS 4.3 is not equivalent to Safari at least as far as the Javascript performance, is it really correct to make this statement in general? While I know that this is not a javascript test, it seems like it has been made clear that Safari is not merely a wrapper for UIWebView (why that is I won’t try to guess) so I am not entirely sure this methodology is actually as fair as might be proposed.

    http://www.theregister.co.uk/2011/03/15/apple_ios_throttles_web_apps_on_home_screen/ for reference…

  • http://www.blaze.io David Horne

    FYI – Official update posted at the top of this post RE: iPhone Nitro JS engine issue.

    • Pat

      Cool – looking forward to the results!

    • Steven Fisher

      Good start. What about every other difference between WebKit and Safari?

      You claimed, after all, that Chrome beat Safari. Since you didn’t test Safari, you’ve got a bit more editing to do.

    • robzr

      David – if you guys are truly comparing the OS, why not use the iPad with the same clock speed as the Nexus S?

      If you’re comparing phones, what is the point of comparing a phone thats been out 2 months to one thats been out 9 months?

      Something smells funny here, either funny or incompetent (and we know how you already put your foot in your mouth with Nitro).

      Rob

  • RichardL

    Ha! Apple gets bitten by their own feature lockdown.

    Karma, baby!

  • http://twitter.com/thedailytechco The Daily Tech

    quite surprising, given the fact that apple had integrated the nitro javascript engine into their mobile safari. not surprised that android was faster, but more surprised by the fact that it was 52% faster

  • nelmat

    I have both an HTC Desire HD and an iPad 1 – the iPad loads websites over 3G 70 – 80% faster than the android phone, on wi-fi, the iPad not only loads faster, but the site is responsive and scrollable significantly faster.

  • EsotericPunk

    If Android’s average browser speed is 2.1 seconds and iPhone’s is 3.2 seconds, this does NOT mean that Android’s browser is 52% faster. 52% faster implies that the Android load speeds are less than half that of the iPhone. Just looking at the numbers, this is clearly not the case.

    The mathematically correct statement is that Android’s browser yielded results that were approximately 35% faster than that of the iPhone, or, better yet, that the iPhone browser loading speeds are approximately 52% slower than Android’s.

    • robzr

      Good point, and I’m sure the editors were smart enough to know that when running the numbers – they can’t be THAT incompetent – I can only imagine it’s intentional. Whats the reason for the bias – sensationalism to drive page hits? PR paid for by someone?

      They also claim they are comparing browsers, when they didn’t run it in browsers and therefore missed the speed benefit of Nitro on 4.3.

      They claim they are comparing the OS, not the hardware, so they choose the 9 month old 777 MHz iPhone 4 instead of the 1 GHz A8 based iPad, which runs the same OS and would be a much more accurate comparison to the 1 GHz A8 based 2 month old Nexus S.

      Somethings off here, sensationalism & dishonesty, incompetence, or marketing. Which one of those is it Blaze? You gonna be honest with us?

      Rob

    • Alvshill

      They made a simple mistake there. I think I’m correct in saying, based on the 2.1 (Android) vs 3.2 (iPhone) load times, that the iPhone takes 52% longer to load webpages. 52% is correct if presented in that context.

  • Hgj Cilbup

    Interestingly, there is no consideration for the spread of results (confidence intervals), also I note there was an effort to remove outliers from the data, which may skew results. There is also no consideration for what the “document call back” means when implemented in the browser engines ie. does this actually mean the entire page has been rendered and is available for scrolling, or the part of the page being viewed is able to be seen (and the rest of the page can be viewed after rendering later.

    It has been commonly reported that there are differences in the browsers in how much of the page is rendered, with regard to zooming, choppiness of scrolling etc.

    In the real world if the browser loads the page slightly slower, but it means I can scroll around like butter, and zoom effortlessly and that takes slightly longer, I am happy with that.

  • Guest

    You must have checked it, but I could not find that you checked whether the devices got served identical/comparable content. Did I skip that, or is that info missing?

  • Anonymous

    You should test local html pages, or pages on a local lan webserver.

    Too many factors applied when testing over the Internet: isp caching, browser caching, isp dns caching, local router dns intercepcion caching, browser agent identification on the webpage, different html code sent to different browsers, etc

    If you want a real world test, do a real world test. Load each page in each phone, one at a time. Record result with a high speed video camera. Since your custom app is not used in the real world, I don’t know what to think about this.

    • http://www.webpagetest.org/ Patrick Meenan

      Loading local web pages is actually the worst thing you could do. By FAR the network latency for each request usually drives the load times so if you got rid of the latency to the servers it would be just as artificial as running sunspider.

      Interesting timing because MS just put up a blog post about IE9 that drives home the point about the network access being the expensive part of loading web pages: http://blogs.msdn.com/b/ie/archive/2011/03/17/internet-explorer-9-network-performance-improvements.aspx

      I expect the async support in Safari that isn’t available in the embedded browser will have a lot more impact than the JS engine. Would also be nice to know if the Android tests also loaded flash which would make the difference even more impressive.

      • Anonymous

        But… What are we benchmarking here?

        This article is about what? The platform official browser, Telecom, ISP, website mobile support, … ?

    • http://snafuracer.blogspot.com/ snafuracer

      True. Real world test would involve less variables, like current network saturation. But I agree on the Java Script, because the mobile web pages don’t usually carry much JS, it’s rather plain HTML, as plain as possible.

  • ARS2600

    Bottom line, we knew all along, Android will beat iPhone very soon. You confirmed it with this article. Thanks much. Android Rules the world. iPhone sucks.

  • Anonymous

    and my htc with 4g plays flash, and its very fast.

  • http://www.facebook.com/profile.php?id=594289803 Jerry Ballard

    Sorry, but if you guys really want to establish some sort of credibility (particularly given your ‘startup’ status), you need to start with a retraction and apology.
    Its clear that you did not test what you were claiming to test, and were testing sandboxed ‘kit’ implementations against each other. That comparison may have been interesting in its own right, but your article is not an honest effort.

  • http://www.facebook.com/paul.readwin Paul Readwin

    My only question is, who cares? Does anyone think that the “speed” of a mobile browser will make or break a sale of a smart phone? I am curious as too what the point is.

  • You liars !

    You’re just fucking bastards. You did know from the very beginning your so called real-world testing was flawed. You just tried to fool your readers. You certainly doesn’t deserve any trust more. For whatever paper to come. You are just liars. You killed the reader trust that gives you your daily food. Too bad.

  • http://www.facebook.com/paul.readwin Paul Readwin

    Actually I do have a theory now after reading all the posts. This isn’t about an iPhone vs. Android showdown. It seems to be a great way to get some attention to a startup that has an “application that automates front end performance optimization. Which they developed as a testing tool to further their expertise in optimizing mobile devices”

    Good work, because it worked :-)

    • Anonymous

      No company ever releases anything on their blog / site unless it is to drive traffic to the site. That’s a no brainer really, otherwise why spend time even writing it? If you’re trying to sell a product, you need eyeballs on your shop window. The more eyeballs, the more sales.

      I’m certainly not suggesting “you personally”, but there is a pretty interesting almost socialist aspect to the web these days where doing something for a profit is seen as a bad thing unless everyone on the web can benefit (preferably for free). The only reason that the big guys like Google get away with mass profiteering it is that the users don’t pay; companies pay instead.

      Regardless of the debate over results (I personally don’t care, I own a Galaxy S and an S2, an iPhone 3GS and 4, an iPad2, a MacBook Pro, a PC, a MacBook Air, 5 more PC laptops etc etc – I’m a developer so I need all that stuff), the fact is that it got you, and me, and loads of other people to click a link, and now some of us know who / what Blaze is.

      Oh, and Android is better from a developer point-of-view. Speed is not as important as the behaviour of the rights-holder. Developing on iOS is fun and easy but you have to contend with annoying and sometimes draconian lock-down that Apple decrees from on high (such as the much touted issue behind this comparison – devs cannot access nitro – how stupid!). Developing on Android sucks donkey-balls and you have to use sucky Java but at least we can do what we want!

      It actually seems like suggesting anything Apple is not as good as product x-y-z is like insulting someone’s religion (which I also don’t hold as wrong). Apple, like all religion is just a made up story selling you something. In this case, it just happens to make your life easier, look pretty and solve some every-day problems.

  • Theanimaster

    Ugh. iOS fanbois complaining about a few milliseconds between the Safari Application and UIWebView (which, I still don’t get) is not the case here.

    The case is — so WHAT if Android is a bit faster? Most people will not notice it even if it were twice as fast. Most people will attribute the slowness to the carrier/DSL speed, NOT the device. Boo-yah b*tch.

    The case is… iOS STILL has it right — easier to find apps, easier to PURCHASE apps (which goes against software pirates)… easier to develop apps — easier for app developers to earn REAL MONEY for their apps.

    So while fandroids want us to look the other way, we’re still ahead of their camp in terms of what we really want to use our devices for — and not some dumb geek numbers.

    • robzr

      This startup Blaze has gotten a lot of publicity out of this – who else reading these comments would have even heard of this company until this came out?

      But it’s a sleazy tactic to ride the coattails of successful companies with flawed benchmarks and misleading headlines.

      The CPU on the Nexus S is 1 GHz vs 777 MHz for the iPhone 4 – so what are we testing, the software, as they purport, or the CPU? If we’re testing the software, why not conduct a fair test on a 1 GHz iPad running iOS? If they want to test latest hardware – why don’t they reach for the dual core A9 based 1 GHz iPad 2?

      If we’re benchmarking hardware, then this is just a ridiculous test. A phone 3/4 of the way through it’s product cycle (iPhone 4 has been out 9 months and will be refreshed in 3) vs. the Nexus S which just came out 2 months ago. Whats the point of that benchmark? No Apple fanboy in their right mind would dispute a 1 GHz CPU of the same architecture would not do something CPU intensive like rendering webpages faster than something running at 777 MHz.

      The headline states “whose browser is faster” – yet they didn’t test the actual browser.

      The byline states “Android 52% faster” yet by their own numbers it’s 35% faster, as noted by EsotericPunk.

      Whoever ran these test is dishonest, incompetent, or both. It will be interesting to see what they publicly acknowledge, if they have the balls to do so.

      Rob

  • Sriranganathan

    Was it performed using seperate wi fi connection. If not, it may reduce the performance.

  • Kentaolsson

    Such an incredible crap talk about Android against Safari! Your test is really misleading – not usual when some people out there in the PC world can´t find their lollipop. Amongst many of us iPhone users it´s a fact that you can´t take advantage of the Nitro engine in Safari when you are browsing through another app. Why should I use another app to browse on Internet? Just start you Safari and there you are. If you compare two or three different Android phones with iPhone beside each other browsing on internet within the Safari application itself, the truth is that Safari on my iPhone4 with iOS4.3 seems to be as fast as on any other androidphone. Why spend 4711 pages and God knows how many hours trying to convince people how bad Safari is!? We iPhone users (and lot´s of Apple users) don´t give a damn about all those freaks yelling against Apple about some useless technical detail. If Apple screwed up some tiny little function in their software, they soon fix it. As usual! And in the mean time 99 percent of Apples costumers are satisfied with their products!! Most of them even with Safari.
    Kenneth

  • j0hnqd03

    If you don’t know how to perform proper software testing, you shouldn’t be doing it, and you especially shouldn’t be posting misinformation online. I wonder who’s payroll you’re on?

  • NeverUsingBlazeProducts

    This is the reason why I will *never* use Blaze products. They just proved themselves as NOT being the expert on development with webkit and the embedded browsing products. But, instead, they tried to get their 15 minutes of fame and blew it. If you try and speak intelligently on a subject, you might want to know a little bit about it first. If you had researched it, you would have seen that Apple *did* make those statements and it was a known fact that the embedded kit had not been updated. Here would have been an easy test for you to do before embarrassing yourselves and your company: Test your embedded browser against Safari’s browser. Oh wait, that wouldn’t have gotten you your 15 minutes would it? Lame move in my opinion.

    In fact, it still says this line in your false report: “Android’s Chrome beat iPhone’s Safari by loading 84% of the websites faster” that is completely false as you stated you ran this as an embedded browser test. Safari is NOT an embedded browser! You never tested Safari! Only Blaze! You guys are so ridiculous it’s not even funny.

  • http://snafuracer.blogspot.com/ snafuracer

    The whole research is flawed from the very beginning, nice PR noise though. I think you got a lot media attention and many clicks … I mean, really, get a grip – who cares about 0.2 seconds slower on a decent phone ?

  • Xilly

    I love my Android, I don’t particularly like Apple, but I still have a hard time trusting the word of someone whose grasp of percentage calculation leads them to claim 2.1 seconds is 52% faster than 3.2 seconds. It’s 34.3% faster. On the other hand, 3.2 seconds is 52% slower than 2.1 seconds.

  • Franky88

    I don’t mind what API what app on the iPhone is using. If it is slower, it is slower.

    The rest is up to Apple, not me and not event to the developer.

  • http://www.blaze.io David Horne

    Blaze follow up post to iPhone vs. Android study. Lessons learned: http://www.blaze.io/business/android-vs-iphone-study-lessons-learned/

  • iPhone user 644786433422

    This whole “test” is crock of shit anyone can say they “tested” this and that so many times and blah blah blah and make a crap diagram saying this or that but I know that when I’m at my job android users ask me to look something up cause their android is too damn confusing and pics on sites are so grainy at zoomed in views that they say you look it up for me plus I have unpersuasively switch them from there crap Droids to iPhones they said they only got there Droids cause of the cheap deals that came with there cheap phones

  • Rfish117

    This IS AN ABSOLUTE LIE.

    I have had MANY Android and have gone through all the iphones. I can tell you that the Android IS ABSOLUTELY NOT AS FAST AS THE IPHONE. Whoever wrote this has absolutely NO CLUE what they are talking about. I WISH my android was faster so I could keep it for the voice texting, the better navigation and the other great features. However, it simply is not. Stop f#@$ing lying.

    Android is also buggy with webpages- scrolling is NOT smooth- but voice search is great.

  • http://www.mangaradja.com Ro3first

    I have both and try to open my site:
    http://www.mangaradja.com
    I feel no different the fastest
    but through PDA windows mobile: ie, its extremely slow ,
    no tool to test it cause you can feel it

  • Aammar Ahmed

    nice information

  • Mohd Ismail98

    The comparison is really outstanding and gave me the insight of the two devices. I thank all the people who have put their precious time and effort to provide the info. It’s really useful. I know you guys will receive negative responses as well, but that shouldn’t be a hindrance to the future efforts. The negative responses will mostly be from people addicted to APPLE (it doesnt mean that i havent used APPLE device). Anyways keep it up.

  • http://twitter.com/jokeaday2011 Markus

    Yay, there is a new Iphone, but with the same old iOS, booo!. The phone can be great, but the OS is not.
    http://www.high-on-it.co.za/2011/08/ten-reasons-why-android-is-superior-to-the-iphone/

  • Generic Person

    Oh and what of you are using 4G?  100% of the time the Android phone was faster,  Why?  Because there is no 4G iPhone (oops, did not mentioned that).

    Why doesn’t anyone mention that?  Oh that is right because you can only be unfair to anyone except Apple.
    Who made that rule?

  • Anonymous

    Go and check every single kind of comparison for Apple Iphone made against an Android phone on every test,be it bench marking,usability,hardware specs,browsing speeds and,alot more kind….and apple fan boys comments of dissatisfaction made it more obvious they have been left behind and they simply couldn’t swallow the truth…..the bitter truth of their expensive I phone is not worth it.

  • http://www.register-domainname.in Register domain India

    Good idea. I used to be staying constantly on this blog and I am inspired! Very much useful information specially the last part which you have posted. I take care of such information much. I used to be seeking this particular information for a long time. Now only I have got it. Thanks for sharing.

  • http://twitter.com/d60pc d60pc

    iPhone with iOS and Android is becoming the most trending smartphone these day. One of the advantage of iPhone is the elegant looks on the hardware and the operating system used, the iOS. iOS not only having a cool features, it also have a good and comfortable interface to be used, but the lackness of iPhone is on the price, it’s pretty expensive for people who does not earn so much. While Android is one of the smartphone with Android operating system as Open Source which provided with lot of free application and games for the users. 

    iPhone with iOS and Android is becoming the most trending smartphone these day. One of the advantage of iPhone is the elegant looks on the hardware and the operating system used, the iOS. iOS not only having a cool features, it also have a good and comfortable interface to be used, but the lackness of iPhone is on the price, it’s pretty expensive for people who does not earn so much. While Android is one of the smartphone with Android operating system as Open Source which provided with lot of free application and games for the users. : http://blogspc.com/ios-on-android-phone-application-to-change-the-interface-of-android-to-iphone-like/3500

    • Scorpineo

      Elegant design?  You mean this elegant design?

      http://en.red-dot.org/2791.html?cHash=de5e1220d7f00b294bd00b01cdcc541b&detail=7398 

      Does that look familiar?  Did Garmin steal the design of the iPhone 4?  No, not unless they own a time machine.  Their design for the Garmin Nuvi 2790T won a design award at least 6mos before the iPhone was even announced! Yes, it won a design award, is it any wonder that the iPhone 4 looks pretty? Apple probably figured that if Garmin makes a stink, they’ll just pay them off, like they paid off others when they stole the names: iPod, iPhone and iPad.

      Yet another example of Apple stealing ideas from others and patenting or claiming them as their own. It’s their corporate mantra. Steal first, pay off or fight in court later…that is, if you get caught.

  • Sam

    apple is awesome

    • http://www.facebook.com/NickgurlPA Stephanie Broadnix Mitchell

      No It isn’t! Android will rule, and iShit will struggle or become obselete

  • vamp

    Wow a lot of butthurt Apple fanboys here. I have both the iPhone 4s and galaxy s2 and s3 and both work much better than the iPhone while browsing.I had more than at least 10 cases of safari just crashing. While people keep saying iPhone just works…most people I know went from ios to Android never would want to go back.