We’ve recently started using RescueTime in the company. I used it personally for a couple of weeks just out of curiosity and found that it was legitimately awesome, so I rolled it out for the whole TopHatMonocle empire (all 6 of us.) It’s been a fun experience.

First off, RescueTime neatly falls into the category of data porn. You get to see so much cool data that you just didn’t have access to before that it’s almost overwhelming. I spent a good hour just exploring all the crazy graphs I could generate by selecting different time intervals and categories – “I spent 12 minutes yesterday on Hacker News… hmmm… I thought it would have been more than that.”

Second, it confirmed my suspicion that I don’t get any work done during the day (i.e. 9-5.) Not to say that I don’t get things done – it’s just that everything I do falls into the vague category of management, rather than things like coding or designing. It’s almost hilarious looking at the graph above, I basically just start to ramp up around 4pm and then dip to zero productivity at around 7am with maybe an hour total work done during the day.

Another side benefit of RescueTime is that people start to compete on who can have more +2 productive time during the day. It becomes a kind of badge of honor to have x hours of productive time. Of course ideally you would compete on actual work done, but this is about as good a proxy for that as you can get – and previously there was just no way to access this data at all.

As an advertiser:

We put in $100 worth of ads, just to test the waters. I wanted to see if it was worthwhile to run a more serious campaign. The first thing that pissed me off was that admob suggests a bid price about 7x what you should actually be bidding. The typical bid is $0.03, but the default suggested bid in the form is $0.20. This is a pretty transparent cash grab, and in general a really shitty thing to do.

For our $100 we got 115,000 impressions. Of these we got 400 clicks (a 0.35% conversion rate.) For those 400 clicks we got precisely 0 sales of our $0.99 app. Based on reading about the experiences of other people using admob, this can be explained by the fact that even that 0.35% conversion rate we got was almost entirely made up of accidental clicks. Other reports suggest that the download rate for free apps is about 13%.

So the conclusion we’ve reached is that if you’re advertising a paid app – don’t. If you’re advertising a free app you’ll be paying about $0.25 per download. To get to the top 100 in the US store for a free app will cost you about $10,000-$20,000. I’m not sure what you’re going to do once you’re there – sell ads? good luck with that. If you’re advertising the lite version of your paid app, the conversion rates for lite-to-pay version of your app is about 1% (usually less – around 0.25%) so you’re certainly not going to make your money back there.

As a publisher:

Admob pays out about $0.01 – 0.03 per click. Yeah.

With a conversion rate of 0.5% (which is above average) you would need about 1,000,000 impressions per day to get revenues of $100 a day. That would require an install base of about 5-10x that. So in other words, unless about 25% of the entire iPhone + iPod market has installed your app, you couldn’t possibly make enough to support even one developer at a shitty salary.

The exciting conclusion:

There is no reason at all to use admob, unless you really want that $3 a day in revenues for a popular app. We’re going to implement a part of our app on a mobile site and use google adsense. It couldn’t possibly give worse results than admob.

Here’s a rule of thumb I’ve come up with:

if( self->studio_type == big_name_studio )
	self->make_app_store_apps();
else
	self->do_something_useful();

Oh wait, what am I doing? that obviously should be done in Cocoa Objective-C

NSString *myStudioNameString = [[NSStudio currentStudio] studioNameStringForCurrentStudio];
NSStudioCompare *myStudioComparison = [[NSStudioCompare alloc] initWithCompasionType:kCFStudioComparisonDefault withStudioType: kCFStudioComparisonBigNameStudio];

if( [myStudioComparison compareToStudioWithName:myStudioNameString withObject:nil withTarget:nil withDelegate:nil withRandomBullshit:nil, nil, nil, NULL, nil] == kCFStudioComparisonResultTrue ) //be sure to put nil, nil, nil, NULL, nil or else inexplicable crash occurs
{
	[self.studioComparisonResultDictionary setObject:[NSNumber numberWithBool:YES] forKey:@"studioComparisonResult"];
	[self.studioComparison setDelegate:self];
	[self.studioComparison commitStudioComparisonCompare withSelector:@selector(studioComparisonDidComplete:)];
}
else
{
	//this has to be performed on the main thread or it does nothing and gives no error!
	[self.studioComparison performSelectorOnMainThread:@selector(studioComparisonResultFalse:::::) withObject:nil withObject:nil withOneMoreObjectJustForFun:NULL withUntilDone:NO];
}

[myStudioComparison release];

//TODO: there is a memory leak somwhere in this function! debugger is useless.

There! now that’s how real programmers code. Of course that’s just the stub. There’s another 300 lines implementing the NSStudioComparisonResultDelegate protocol. Obviously.

In the near future I will be posting about our upcoming project, which will be the focus of the company for the next few years. I’m really excited about it, and I wish I could say more now, but it’s just too early. In the meantime watch this site as we post progress with Hopple and Dante as well as anything else interesting that happens.