It's Rishi

Paul Buchheit, the creator of GMail, and a founder of FriendFeed (which I wrote about earlier) wrote an interesting post describing his philosophy on the development approach of innovative products (typically in startups). I found his thoughts to be very similar to those of my own. My favorite part is:

So what’s the right attitude? Humility. It doesn’t matter how smart and successful and qualified you are, you simply don’t know what you’re doing. The good news is that nobody else does either, though some are foolish enough to think that they do (and that’s why you can beat them).

What is the humble approach to product design? Pay attention. Notice which things are working and which aren’t. Experiment and iterate. Question your assumptions. Remember that you are wrong about a lot of things. Watch for the signals. Lose your technical and design snobbery. Whatever works, works.

What I tell people over and over is that one can be the most accomplished product designer/manager/engineer, but when developing a new product, you are really just making an educated guess about what will resonate with your user. Sometimes what makes so much sense on paper just doesn’t jive with users. In a sense, the design+requirements for the initial product is the hypothesis and the v1.0 of the product is the experiment that tests the hypothesis on users.

What separates the winners from the losers is the analysis of the results, which in the case of web-based products can be efficiently done by looking at specific engagement metrics. This does not just mean pouring through Google Analytics data. Instead, I’ve found it to mean combining the analytics data with database queries that measures key application engagement metrics.

The point is that the development of innovative products must be both rigorous and methodical. Use the standard scientific method. The unknown question is “What do my users want?”. Start with a hypothesis, experiment by testing your products with real users, analyze what worked and what didn’t, modify your hypothesis, test, …

For a couple of years, I’ve had this recurring thought: “What if you could apply the PageRank concept of link popularity to reputation?” In other words, let’s say person A, a marketing guy, vouches for person J as being an awesome Java engineer. That endorsement, while it may well be true, would not mean nearly as much as if person B, an engineer himself who has 10 other engineers vouching for his Java engineering skills, vouches for person J’s Java skills. Well, imagine a giant graph with edges that represent such vouches. By analyzing the graph, one could find the best (essentially the node on the graph with the most incoming Java engineer vouches…again not just quantity of vouches but quality of vouches…sort of a weighted sum) Java engineer.

Well, as a sort of experiment, Paul and Nick were kind enough to help me whip up the Vouch For Me app on Facebook. Add it and start vouching for your friends and get vouches back.

Lookery, an ad network for Facebook and other social networks with developer platforms, announced that, as part of their effort to rapidly grow their inventory, they are offering guaranteed minimum of 12.5 cents CPM for the next three months for an exclusive on the application’s traffic. 12.5 cents CPM at first glance seems awfully low but one must consider the alternatives.

Ad rates, in this case we’re talking in terms of CPM, generally correlate directly with context. Stronger context means a more focused audience. Targeted ads for this audience brings in high rates. For Facebook apps that have pages with real context, it’s likely that they can and should do much better than 12.5 cents. Depending on the content genre, it may take a bit more legwork on the part of the developers to find the right advertisers who will pay, but if the volume is there, it shouldn’t be too hard. However, for apps that have no real context (the majority of apps), 12.5 cents CPM may be about as good as they can get. Facebook itself sells its “flyer” ad space (the left column under the nav links) for not much more than this. In fact, I recently paid about 20-25 cents CPM for an ad and that was when I instructed Facebook to serve my ad only to a very precise demographic (matching only ~40k users out of the 50M+ Facebook userbase). Without such targeting, I would have gotten away with paying much less.

From my own experiences, I’ve found SocialMedia to be quite lucrative. SocialMedia advertisers, primarily developers who are buying installs for their own applications, pay 15 cents and upwards for a click. When SocialMedia first launched, when we published their ads prominently at the top of our canvas pages, we saw eCPM as high as $2 or so. Now it’s less, but still is above 50 cents eCPM. My hunch though is that applications which serve massive page views, particularly a large number of page views per user session, will see decreasing eCPM from SocialMedia. There’s only so many ads that each user is going to click in a session, no matter how many page views in length. Again, this is just a hunch though.

So, if you are one of the application developers that has a massively popular application which offers no meaningful context, then Lookery’s offer is probably pretty attractive. And this type of developer is exactly who Lookery wants in order to achieve their goal of adding a billion page views of inventory a month for the next few months.