Part of the problem might be that solutions that work for US startups (and therefore are assumed to be defacto global solutions) might not work in Africa. The usual promotional tools for a website are to get a blog up, build your followers on Twitter, try to get covered by Techcrunch etc. None of this is going to be anywhere near as effective in Africa. It's probably not cool - but perhaps African (and all developing world) entrepreneurs should look back to see how Web 1.0 sites built audiences - ebay, paypal, god forbid... even Yahoo.
Edit: I think it's taken as a given that this extends to the technical too. While we're all bitching about IE6 and drinking our Ajax goodness from the cup of rails, it's easy to forget that there are guys who should be banging out sites in 800 x 600 optimized for dialup.
There's the pricing factor too... a population with a median income of between 2 and 10% of US median income simply isn't going to buy SaaS subscriptions in comparable volumes at comparable prices to US companies. For much the same reason, these consumers are also worth less in raw money terms to advertisers.
which means a lot less revenue for typical Y-combinator startups in African domestic markets even before transaction costs (quite possibly higher?) are taken into account.
Edit: I think it's taken as a given that this extends to the technical too. While we're all bitching about IE6 and drinking our Ajax goodness from the cup of rails, it's easy to forget that there are guys who should be banging out sites in 800 x 600 optimized for dialup.