The average web page has grown 20% in just six months

The HTTPArchive celebrates its second birthday this month, and it seems fitting to check in and take a quick snapshot of a typical web page. Not surprisingly, pages are bigger. But surprisingly, pages are even bigger than I expected.

Here you can see the growth in average page size over just two years:

Web page growth: November 2010 to November 2012

When I checked last spring, the average page came in at 1042 KB, just over 1 MB. In other words, a typical web page carries a payload that’s 20% bigger than it was just six months ago.

Anecdotally, I’ve noticed an interesting trend over the past few years. At this time of year, pages suffer “holiday bloat” as site owners try to stuff in more images, Flash, and rich media, not to mention a fresh crop of third-party scripts for things like ads, analytics, and trackers. Not unlike us Westerners, pages get fatter throughout November and December, and then, interestingly, trim down a bit when they go on a crash diet in January. At the risk of pushing this analogy over the edge, I’ll add that, like all crash diets, the effects don’t last. Before long, pages start puffing up again.

Pages could hit 2 MB in early 2014

When I wrote about page bloat for GigaOM last May, I projected that, at the then-current rate of growth, the average page would hit 2 MB in 2015. Based on this new finding, we need to upgrade that prediction to early 2014:

Projected web page growth: November 2012 to May 2014

Page bloat in and of itself isn’t news. Page bloat of this magnitude is. This has obvious ramifications for desktop users, but mobile users will be much harder hit.

For more insight into the current state of web pages, I recommend you check out Andrew King’s recent post Average Number of Web Page Objects Breaks 100, as well as Catchpoint’s Holiday Shopping 2012 State of Web Performance.

Related posts:

10 thoughts on “The average web page has grown 20% in just six months

  1. Pingback: Holiday Shopping 2012, State of Web Performance | Web Performance Monitoring and Optimization

  2. Great article Josh. This might be a detail but since I’m responsible for the presentation of the HTTP Archive’s data I wanted to make sure to be clear on the issue of *sample size* and how it affects the absolute value of some of these stats, although the (alarming) trends are still absolutely spot on.

    Here’s the sample size issue: The HTTP Archive is still growing and thus the number of URLs that are analyzed is growing. You can see this if you look at this chart of “URLs Analyzed” over time. Back in Nov 2010 we only analyzed the world’s top 16,656 URLs, whereas in Nov 2012 that number increased to the top 292,999 URLs.

    I believe you’re data comes from looking at the “Total Transfer Size” chart for ALL sites. Since performance best practices (such as keeping the page size small) are better for the most popular websites and get worse as we approach the tail of popular sites, part of the increase you’re seeing might coincide with the change in how many URLs are analyzed. For example, in November 2011 we jumped from 16K to 50K URLs, and the total transfer size also jumped.

    It’s possible that this increase in transfer size could be due to “holiday bloat”. One way that we can eliminate sample size from the equation is to look at “Total Transfer Size” for the TOP 1000 sites. The sample set is fairly constant, so increases in size is more directly due to the website adding more & bigger resources. For the Top 1000 URLs the total transfer size grew from 626K (Nov 2010) to 784K (Nov 2011) to 1,114K (Nov 2012). The first year saw 26% growth, while this last year grew 42%. That’s an alarming change in the slope for these top sites.

    I’ll apologize for the charts being hard to read and manipulate. It should be possible to hover over the line plots to see values more clearly. And unfortunately the sample size will continue to grow in our quest to hit 1M URLs. One way I hope to address this is to add Top 10K and Top 100K choices to the existing Top 100 and Top 1000 choices. That will allow for viewing trends over larger populations without concern about sample set.

    This was long. Probably not many folks need to worry about this level of detail. But for those who sweat the details I wanted to put it out there.

  3. I wonder what would happen if you corrected for in-post images vs structural page weight? My first inclination is to think the results here are heavily weighted in favor of image-heavy blogs and other “look at me” style sites. If you were to treat product images as post content, I suspect we’d see a further reduction in average page size again.

    I guess my point here is this – I feel like we’re looking at the results of what laypeople do with their websites and not what the designers are putting out there. Most reputable designers that I know practically obsess over their results at webpagetest.org and Pingdom.

  4. Pingback: 2013 predictions: The average web page will hit 2MB, Android will pull ahead of iOS for good, and your smartphone will get infected with a virus

  5. Pingback: Mobile Browsers Help Users Avoid Bloated Webpages | PJExploration

  6. Pingback: Mobile Browsers Help Users Avoid Bloated Webpages | 5 For Business

  7. Pingback: Mobile Browsers Help Users Avoid Bloated Webpages | Android Developers

  8. Pingback: Mobile Browsers Help Users Avoid Bloated Webpages

  9. Pingback: | Web Design Florida Organization

  10. Pingback: Blue Note Tech Blog » Mobile Browsers Help Users Avoid Bloated Webpages

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>