Phillip Kruger argues that responsive design shouldn’t be used in Africa in a Memeburn article called Why Google might just be wrong about responsive design in Africa. He lays out his argument in two parts. First:

Responsive design only works on smartphones. So by default you are already ignoring 80+% of users in Africa. You are also reaching the 20% of users that possibly have internet access at home or work.

This is an argument I see a lot, but it’s valid only in the context of target audience and use cases. If the target market for your service primarily uses low-end phones, then by all means don’t bother with responsive design. But let’s say you’re building a site to order take-out food and deliver in major cities, the situation changes dramatically. Now you’re most likely looking at a target market that sits squarely in the 20% of people who have smartphone access (and who don’t want to get off their couches and walk to their PCs to place an order).

This is why personas, scenarios, and use cases are so important. If you’re building a service for ALL THE PEOPLE (which isn’t advisable), then averages are appropriate. And those averages will rightly guide you to focusing on the 80% of people who do not have smartphones. But in most cases, the analytics that matter are not the averages of all users, but the specifics of the market you’re going after. Don’t dismiss responsive design in Africa because of averages. Dismiss it if it doesn’t make sense for your target market.

Phillip continues:

Responsive design is not lightweight. When using responsive design, the size of the download to the browser is still very big (in fact it’s very similar to what the webpage would be). All the HTML is still being downloaded (even parts that are hidden on a smartphone if you use media queries to set display:none in your CSS). Sure, you can have rules to download separate images for separate display sizes and that should help a bit.

Tim Lind wrote a comment on the article that addresses the problem with this logic:

Responsive design does not mean you can’t do server side optimisations. In fact it can help you to do these, and encourages a more efficient design process. You can design the content for the feature phones, and responsive media queries allows you to upgrade the design with a single stylesheet file for the smartphone or desktop (which server side optimisation could exclude).

One of the many good things about responsive design is that it forces designers and developers to spend a lot of time on optimisation to ensure light, fast pages. This is just good practice for web development in general — not just for mobile. Page bloat is a huge problem, with the average web page now being more than 1 MB big.

Page speed optimisation is just good web citizenship, and it should be a requirement regardless of whether or not a responsive design approach is taken. The other point to remember is that mobile networks already do a lot of compression on served images (see How should we handle responsive images?).

What worries me about this debate is that there appears to be no room for nuance. Responsive design is either the answer to all of Africa’s problems, or we shouldn’t do it at all. But as with most things, the appropriate approach is to say “it depends.” A mobile strategy shouldn’t be a decision between a native app or a separate mobile site. A mobile strategy should form part of a larger web strategy, and it needs to include a discussion about the appropriateness of responsive design. It might not be the right thing for your project, but it should be on the table.

I keep reminding myself of Ben Callahan’s statement in The Responsive Dip: “Just because you can’t, doesn’t mean you shouldn’t.” Just because this is a difficult problem that we haven’t quite figured out, doesn’t mean we should throw it away and go back to how we’ve always done things. What we need to do now is push through and find elegant ways to apply responsive design in Africa. Where it makes sense, of course.

Update 2012/11/26: Phillip responded to all the feedback on his Memeburn post. See Google might be wrong – part 2. It’s good to get additional clarification on the Google talk that formed the backdrop for his original post. This isn’t the last discussion we’ll have about RWD in Africa, and that’s a good thing. We need to figure this out…