User Acquisition and Retention

There is a direct relationship between how fast web pages load and conversion rates. If pages don’t load fast enough, users will abandon whatever they are doing, including buying your products. On the other hand, the slimmer the data you pack into your product page, the faster it will load, and the more likely users will buy and return in the future.

The average statistics show that one in four users will abandon a mobile page load that takes longer than four seconds. It’s worth testing your pages, because this could mean massive impact on the bottom line, just from page loads. If you need convincing, here are some real-life stories:

• Amazon has shown that for every 100 milliseconds of slowdown, it experienced a 1% drop in revenue. Or, to take a different turn on stats, a giant, such as Amazon, would suffer an annual loss of $1.6 billion if its pages loaded just 1 second slower. On the other hand, Amazon increased revenue by 1% for every 100 milliseconds of improved page load speed.

• Walmart’s latest reports show that for every 1 second of page load improvement, it experienced up to a 2% increase in conversions. For every 100 milliseconds of improvement, it grew incremental revenue by up to 1%.

• Shopzilla sped up its average page load time from 6 seconds to 1.2 seconds and increased revenue by 12%, as well as page views by 25%.

• Small sites, such as AutoAnything, cut their load time in half and saw revenue grow by 13%.

• President Obama’s 2012 presidential campaign based its entire fundraising success on making its website load instantly.

In addition, better site reviews, more word-of-mouth downloads, and better retention can probably save your marketing departments lots of money.

Running Costs

The content of your websites has to be stored somewhere, and that somewhere is usually a bunch of hard drives in a big fat cloud. Hard drives cost money. Shipping data to and from the cloud costs money. Renting (or building and running) data center space and bandwidth costs money. And even though cloud technology is normalizing and costs are coming down significantly, bandwidth and storage remain challenging financial problems for big companies.

In 2015, Netflix announced that it was going to begin modifying the compression technology for its video streams; adjusting which algorithm was used on what video depending on the noisiness of the content itself. This move was projected to save the company massive amounts in terms of bandwidth costs, and allowed it to accommodate to the performance of specific devices.

Also in 2015, Facebook published details on how it serves preview images, which are only 200 bytes each. Considering how many images the social media network sends around on a daily basis, this was a huge win for them, especially for users on 2G devices.

You don’t need an advanced degree in mathematics to understand this one: smaller data results in smaller outbound costs, less inbound costs, and fewer storage costs.

Planning Ahead

Web pages are getting bigger. An independent website, HTTPArchive, has been downloading, cataloging, and gathering statistics on the top 1,000 web pages since about 2011. One of the interesting stats gathered during analysis is the average size of these websites. This statistic sums the bytes required to display the page information, including JavaScript, HTML, CSS, JPG, and video. This page size reportedly grew by 24% in 2013, landing at about 1.5 MB per page, and in 2015, it hit 2 MB in size.

One reason why websites are growing larger is because they contain more images, and those images are increasing in size. In addition, there is more code as websites are becoming more complex overall. This problem is becoming so bad for users on 2G, that big companies like Google launched a new framework called Accelerated Mobile Pages, which specifically targets sites to reduce their dependency graph, image sizes, and basically tries to offer leaner, faster content to users on bad bandwidth.

The Next Five Billion Users

There are 7.4 billion humans on the planet right now. About 2 billion of them are currently connected to the Internet. Most of the rest live in countries that are rapidly expanding their connectivity. This means that your largest growth potential is in the emerging markets of Asia and Africa. The advancement in mobile computing technology over the last decade means that the next billion humans will be coming online, for the first time, from a mobile phone rather than a desktop or laptop device.

In their book, New Digital Age (Penguin Random House), authors Eric Schmidt and Jared Cohen lay out the topic well:

There are already more than 650 million mobile phone users in Africa, and close to 3 billion across Asia. The majority of these people are using basic-feature phones—voice calls and text messages only—because the cost of data service in their countries is often prohibitively expensive, so that even those who can buy web-enabled phones or smartphones cannot use them affordably. This will change, and when it does, the smartphone revolution will profoundly benefit these populations.

Mobile Networks

Building networks for these high-saturation-potential countries will not be cheap. If you consider that Verizon had to shell out $50 billion to simply upgrade its network to 4G, the cost of building an entirely new network for such a massive population must be astronomical. Such costs will often be inflated due to government entanglement (as is common for government and telecom companies), and this generally will result in passing off all costs to the end customer.

The world has been seeing great improvement in network speeds over the past few years. However, it’s important to see how this improvement is not uniform in terms of numbers, or geolocation. Google Analytics has a set of fantastic charts showing the trends in connectivity, worldwide. It’s easy to see that the idea of improvement is not homogeneous. For instance, China saw an 8% increase in median page load time for desktop computers (things got slower), whereas their mobile performance time decreased 33% (things got faster), but still landing at >3.5 seconds load time. This is a pretty big number, considering 42% of their 1.53 billion population is online.

The short form is this: mobile networks will continue to grind their way to increase speed, slowly, unevenly, and at great expense. If you’re waiting for the mobile web to suddenly get faster, you might need to find a more comfortable chair to wait around in.

For example, a 2G network has ~0.021 MB/sec transfer speed, whereas GZIP can compress 61 MB/sec. Even a reduction of 10 times in GZIP speed would still access a single megabyte faster than the network could transfer it. Colt’s analysis of these data points suggests that it would be cheaper to invest in a better compressing-slower decompression codec than it would be to invest millions of dollars upgrading the network hardware.