Apps are everywhere. You're bombarded with calls to download apps from your supermarket shopping cart to your local newscaster closing out the 6:00 show. It's gone so far, that Chris Anderson recently proclaimed "The Web is Dead" and that simple, sleek services (aka apps) are the future of the internet. I wouldn't quite go that far (who the hell has even heard of this Chris Anderson dude anyways), but it's clear that the mobile application space is growing at a rate even Groupon would be jealous of. According to a recent study from International Data Corporation, mobile app revenues will surpass $35 billion by 2014.

Here's something I can say with certainty - over the next 5 years, just about everything that can have some type of technical interaction will be turned into an application. It's the growth of an entire industry happening before our eyes. It was just yesterday when it was crawling (the days of Fart Machine) and now it's already telling you that you don't know anything and it wants to hang out with its friends (today - Angry Birds and Flipboard). In case you can't tell, that shitty metaphor was my cute attempt at saying this industry is growing incredibly fast and is bringing to surface some incredible challenges. Most notably - how am I supposed to find the perfect application for X?

Right now, the most popular methods of discovery for applications are the manufacturer/OS supported app marketplaces. By this I mean Android Market for Android, App Store for iOS, and App World for Blackberry (even though they're really a drop in the bucket). In fact, these marketplaces are still so effective that the "holy grail" for most application developers is to be listed as one of the "Most Popular" apps within a given vertical because this placement drives an overwhelming number of downloads. Besides a few editorially chosen apps and some "what are the masses downloading?" methods of organization, these app marketplaces are driven by search. In turn, the likelihood of an application showing up for a specific topic or search term is more so based on the install copy that the developer writes (rather than what the app actually does or how well it does it).

So let's go back to the question I posed earlier - how am I supposed to find the perfect application for X? Using the manufacturer/OS supported app marketplaces, you're going to have a very tough time. They're going to return what's most popular or what matches the install copy the closest.

This is clearly a problem. How do we solve this problem of application search and discovery? There have been a couple companies that have taken a stab at this in different ways. I've divided them up into three primary categories: the depth strategy, the social strategy, and the algorithmic strategy.

GetJar is what I'd call an alternative app marketplace that is pursuing the depth strategy through and through. Sporting several million applications available for download across a very wide range of platforms (everything from old feature phones to newer Android devices), they're clearly the largest application marketplace in the space. The problem is that it sucks. They're completely about quantity over quality (I suppose an argument can be made that they're not even an app discovery mechanism but whatever, I've already written this much). For instance, I performed a search on a pretty generic vertical - "News" - and got back a list of terrible applications with 1-10 downloads and 0 reviews. My gut tells me that the reason this crap shows up first is because people are paying for the placement. If so, that's great for GetJar but terrible for everyone else. Generally speaking, the notion of having the most of something (in any vertical) isn't really a good strategy unless you have a strong means of mobilizing your users/consumers/eyeballs to take action. GetJar doesn't.

I also took a look at two interesting companies that are taking a more social approach to the app discovery issue - Chomp and Mobspot. By using a combination of reviews on major app marketplaces (like the App Store and Android Market) as well as the opinions of your friends, they're able to suggest some interesting applications based on your search terms. Between the two, I found Chomp to be more useful because it was less dependent on my social graph and more so upon their own data and reviews. Mobspot was a little more difficult to navigate. The service is so dependent on your social feed and the opinions of its users that it's hard to get use out of it. As they scale, this could rapidly change though. Overall, I think that a social heavy solution to this strategy could prove to be extremely effective (I can say from experience that most of the apps my friends like, I like too). It's just hard to speak to the efficacy of this right now since I can't really find an example of it that's kick ass.

For an algorithmic strategy, I looked at StumbleUpon and app|esp. I think the thought behind app|esp - it scans your phone's hard drive to see which apps you've downloaded and recommends new applications based on that data - is a good one. However, I was thoroughly unimpressed with the recommendations they gave me. They were extremely generic (apps like Google Goggles and Paypal) or didn't really relate to anything I had, or would want to have, on my phone (I didn't have any shopping related applications but two of the top 10 recommended apps were barcode scanners - does that make any sense?). StumbleUpon works similarly to its web counterpart - you select a few topics and then it scans your phone to see what you've already downloaded. From there, it recommends applications based on a wide range of factors (what you've downloaded, what other users have downloaded, your topics of interest, etc.) and encourages you to rate them with a thumbs up or thumbs down. Over time, it's supposed to get better at recommending applications based on what you've like and haven't liked (a la Netflix). Additionally, if you integrate StumbleUpon with your social graph (your Twitter or Facebook accounts in particular), you can blast out your favorite apps to friends and see which ones your friends are downloading. In a nutshell, StumbleUpon was my favorite. The apps they showed me were all pretty good (a lot of them didn't necessarily interest me but they seemed well executed) and the fact that they rely heavily on their own data (as opposed to my social graph) didn't require me to have a ton of nerdy, StumbleUpon-using friends in order to get some value out of the product. But, in terms of it's ability to help me find an application that performs a specific function, StumbleUpon was pretty terrible.

What sticks out to me the most after running through all these services is that the most effective concepts are those that straddle the line between social and algorithmic (in the same way the the current web ecosystem is forcing users to straddle a line between conventional search and social search). Anything too algorithmic doesn't provide very relevant recommendations and anything too social usually didn't have enough data. But, app search and discovery is going to need to progress beyond the current social and algorithmic strategies if its ever going to be very helpful in answering the question - how am I supposed to find the perfect application for X? It's funny to think about this problem in the context of the web where search is so mature that the only improvements that can be made upon it are to make it faster and more real time. But with apps, we have a serious problem on our hands. The person who helps users find the perfect application for expensing travel, getting your baby to sleep, figuring out the right painting to hang on your pale green wall, or anything else you can think of is going to cash in big. The question remains - how do you determine the "perfect" application for a specific topic? Should it incorporate what your friends say, what the crowds download, the color of the icon, the content of the source code, or the number of install page views between midnight and 1 am? It's probably going to be all those things and more. But let's be honest, if I knew the answer to that question, I wouldn't tell you...