This post about competitive dynamics has been stewing in my mind for months now and it’s still a work in progress.  At its heart is a framework for thinking about a common type of tech company:  the aggregator.   The aggregator takes disparate items, gathers them, and presents them as a unified front.

Aggregators can exist for both content and also products/services and there’s thousands of examples of them across every category:  Google News (news content), OpenTable (restaurants), Expedia (airlines and hotels), Lendingtree (loans), SeamlessWeb (restaurant delivery), Digg (web content), Servicemagic (service contractors), Zocdoc (doctors), Admob (mobile ad units), AdWhirl (mobile ad networks), Pontiflex (marketing leads), GymTicket (gyms).

Convenience is often the key value these aggregators offer: a one-stop stop for customers to find what they’re looking for without going to ten different places.  The ability to compare items is also important. 

In almost every case there’s a interesting tension between these ‘aggregators’ and their ‘constituents.’  Let’s consider Google News.  Google news is increasingly the starting point for people looking for news on the internet. Newspapers hate that Google News is scraping their content and eroding their brand value — but at the same time, Google News drives a significant proportion of their web traffic.  They’d be stupid not to want that. As a member of an aggregator, they’re ensuring they get web traffic.  Unfortunately they’re helping build the Google News brand rather than their own.

Are they shooting themselves in the foot?

This issue arose in my post about Opentable.  One commenter wrote restaurants participating in OT, build the OpenTable brand rather than the restaurant’s own brand. It’s true!  But what can done?

Once established, the aggregator has the upper hand. All the individual entities/constituents act in their own self-interest and therefore will remain part of the network.  No single constituent can defect without suffering harm.  And widespread rebellion / mutiny is unlikely — it’s unlikely that all the restaurants are going to band together and start their own version of OpenTable.  It’s a tragedy of the commons, and the aggregators benefit handsomely from the resulting lock-in network effect.

As an established aggregator, risk can come from only a few places:

1) Competition in the form of another aggregator

2) One or more constituents decide to sidestep you.

#1 is hard to avoid.  #2 is rare, but extremely interesting when it does happen.  One example of this is Southwest Airlines, which isn’t listed on any of the travel booking sites.  Similarly, Admob refused to serve ads through AdWhirl, an ad network aggregator (and when that didn’t work they bought ‘em!)

Occasionally the constituents themselves will ally:  One example is Hulu, a joint venture between NBC, FOX and ABC, which aggregates all their content into a single place.

And once in a blue moon a constituent will creatively embrace aggregation in their attempt to fight the aggregators.  For example, Progressive Auto Insurance proudly shows you the prices of their competitors alongside their own prices.  Fascinating strategy.

The more fractured and crowded the marketplace, the less likely a mutiny or rebellion.  Are the thousands of restaurants on Seamlessweb suddenly going to unite to form their own online ordering system and destroy Seamlessweb?  Not likely.  Are the dozen or so large newspapers going to unite to rally against google news and demand to be de-listed or compensated better?  Absolutely.

As the number of constituents increases, the dependency on any one constituent decreases.  And as an aggregator grows its brand, it becomes extremely difficult for a constituent to break away.  Doing so requires an extremely strong brand and unique offering (like Southwest Airlines) and an alternative sales/delivery channel.

This is most important in the context of a offline company: Consider that Brick and Mortar stores like Walmart are essentially product aggregators.  Shoppers go to Walmart because they know it has a wide selection at great prices.  Suppliers don’t want to miss out on the huge volume that the Walmart sales channel delivers.  The more Walmart grows, the more crucial they become to their suppliers’ businesses.  And the more suppliers they gain, the more crucial they become to consumers.  At the end of the day, Walmart has incredible pricing leverage over its constituent suppliers. There simply aren’t many alternative channels.  Suppliers are trapped.

I’m going to end the post here because it’s already way too long.  But please leave your thoughts and help me push this topic further.  Thank you!

Lions, Tigers, and Spyware on Phones, Oh My!

Mobile spyware is the focus of the tech media’s latest frenzy. It started when a hacker discovered that the Pre sends back location data about users to Palm.  Next, a blogger ‘discovered’ that certain iPhone apps also phone home.  The frenzy came to a head when ReadWriteWeb published Dear iPhone Users: Your Apps are Spying on You.

(from www.flickr.com/photos/gerlos/3119891607/)

This article focused on the NYC-based iPhone Analytics company Pinch Media. The issue? Pinch Media’s software allows developers to learn a lot about their users: Apps with geolocation features can return information about the location of their users. Apps using Facebook Connect can even return demographic information (gender and age) about their users.

Of course, there’s no personally identifiable information here. It’s all aggregate anonymous information — and this has been Pinch Media’s response to the issue. Tracking anonymous information for benign purposes is analytics — not spyware. At the end of the day, developers simply don’t know all that much about their individual users. It’s not like they can identify them by name, right? RIGHT?

Well, um, on that note… we know the full name and location of each and every Android user with our app.

How?!  Did we build in some sneaky spyware into Exit Strategy NYC?

Nope.  Google tells us. This information is part of the Google checkout process behind android app purchases.  Each app download contains the full name of the user:

Clicking on the order number reveals a more detailed page containing the billing city and zip code of the user:

Creepy?  Absolutely.  A google/facebook/linkedin search can reveal incredibly detailed information about every android user with our app. Furthermore, this information is pushed on us — I certainly didn’t choose to see this detail about our users!

Seeing this level of user information displayed was extremely alarming at first. But when you think about it, it’s really not that surprising. Google Android purchases are processed through Google Checkout — the same system that applies to e-commerce transactions.  Certainly I would need to know my customers’ personal information if I were shipping a physical product.  Should digital purchases be any different?

…or “So You Think You’ve Got a Million Dollar App Idea”

(this piece also appeared on Silicon Alley Insider)

As a number obsessed techie and ex-management consultant, market sizing and research were a big part of my launch preparations for Exit Strategy NYC. Since launch, I’ve received many questions from people struggling to estimate the market for their iPhone app ideas.

I’ve put together this document as a guide for entrepreneurs considering developing an app. Below, I’ve compiled some up-to-date numbers about Apple devices. I’ve also laid out a framework for estimating what kind of sales can be expected from a paid app.

The Basic Facts

• 45 million iPhone and iPod Touch devices [Apple Earnings Announcement]
• 54% of iPhone and iPod Touch users are in the US as of June 2009 [Admob Mobile Metrics Report]
• The iPhone comprises 68% of worldwide iPhone OS devices and the iPod Touch makes up the other 32% [Admob Mobile Metrics Report]
• Only 75% of users actually download apps [Pinch Media]
  • The most frequently downloaded free apps reach approximately 30% of devices [comScore]
  • The most frequently downloaded paid apps reach approximately 3% of devices [My calculations - explained later]

Right off the bat, there’s a few back of the envelope calculations to make: 54% of the 45M devices are in the US which means ~ 25M devices. The US has about 300M people.  That means about 8% of the general American population has one of these devices.

How To Use These Numbers

Combine this data with your own numbers about how large of a market your product is addressing. For Exit Strategy NYC, our addressable market consists of all subway riding New Yorkers. In 2008, there were about 5M weekday riders and about 3M Saturday riders [MTA's ridership numbers]. The Saturday number is the more relevant one as it better captures subway usage by NYC residents rather than regional commuters. Neither number counts unique riders though, and given that there are 8M residents of NYC our addressable market size is probably somewhere in between these numbers. Let’s say 6M subway riders.

New Yorkers probably skew more techie than average, so let’s assume 10% (rather than 8%) have an Apple device. Also, Exit Strategy NYC works on both iPhone and iPod Touch devices. If your app requires phone/gps/camera/internet to work well, exclude iPod Touch users from your calculations.

How many Apple device toting subway riding New Yorkers are there?  Well 6M subway riders with 10% penetration = 600,000 potential users.

“But How Many People Will Actually Buy My App!?”

Entrepreneurs are optimists by nature, and it’s tempting to think that 100% of people will buy your product. After all, your product is awesome, right? But reality is a quite different story. In fact, only about 3% of users have purchased the most popular paid apps. To determine that number, I used sales figures from one of the all time best selling paid apps, Firemint’s Flight Control game. According to Firemint’s Alexandra Peters, sales to date have been 1.4 million. As a percentage of the 45M Apple devices, this is ~ 3%.

You should expect a similar upper bound of 3% to apply to whatever market vertical you’re addressing. Of course it’s possible that your app meets some crucial compelling need and therefore achieves a higher penetration rate in your vertical. But don’t count on it — it’s equally possible that your app gets lost in the noise and can’t get traction. Flight control has held a constant spot on the top paid app list for months now. Few others have this advantage.

Realistic Unit Sales Calculations

Returning to the Exit Strategy NYC figures, we knew that if we had an effective marketing and press strategy, we could probably achieve something close to this 3% penetration figure — perhaps higher as many New Yorkers are very passionate about the subway (see? there’s that ever-present entrepreneurial optimism!). 3% of the 600,000 subway riding devices would mean 18,000 unit sales. Does this translate to $18,000 total sales? Our maximum penetration figure was based on a 99c app, but what effect would Exit Strategy NYC’s $1.99 or $2.99 price point have on total sales figures?

Factoring in price into market sizing is difficult. Based on our own informal market surveying, we estimated that the most profitable price point would be $2.99 or $1.99. Around 75% of people willing to pay 99c would also pay $1.99 or $2.99. So 75% of 18,000 units at those prices works out to an ballpark range of around $27k – $40k. Like all software, the app’s unit costs are zero, it’s important to focus on maximizing dollar sales rather than unit sales.

A Growing Platform

One thing to remember is that the user base for apps is growing by leaps and bounds. In their latest quarter, Apple sold 5 million iPhones and 3 million iPod Touches. This means that the potential market for an app grew by more than 20% in only 3 months!

Non-Apple Platforms

One last thing to note: The iPhone certainly dominates headlines, but it’s not the only game in town. In fact, Blackberry outsells the iPhone every day. And in a town dominated by Wall Street, it seems like everyone and their mother owns a Blackberry. Realizing this, we carefully designed Exit Strategy NYC to be easily portable across different mobile platforms. Our app is available for iPhone, iPod Touch, Blackberry Bold, Curve, and Storm, Android Phones, and even as an e-book on Amazon Kindle.  Combined, the Exit Strategy App reaches a significant portion of New Yorkers.

But are device sales a good indicator of a platform’s expected app sales?

Stay tuned to Back of the Envelope to find out.

I don’t usually enjoy popping balloons, but there’s way too much hot air going around these days.  It’s time that somebody tells the truth about the current state of iPhone app advertising.  I hear too often from would-be iPhone app developers that making big bucks with ad supported apps is easy: Just stick in some $30 CPM ads, sit back, and watch the money roll in!

To understand why naive first-time developers have this mindset, you only have to turn to the figures being tossed out by the major iPhone ad networks.  Last summer, Admob was talking about $30 CPM brand ads and calling that “low end.”  Similarly, Medialets talks about their Dockers ads which paid in the $20-30 CPM range.  Even in today’s tough advertising market, Admob company continues to cite rather high “$12 to $14 average CPM” figures.

Ready for the brutal truth?  Most iPhone ads networks today pay around $0.50 CPM. In case you don’t know how to digest that statement, I’ve rewritten here in plain english: 1000 people have to look at your application’s ad just to earn you a measly 50 cents.  What about those $30 CPM figures?  They’re just marketing fluff.

If you want the truth, ask the folks on the front line: actual developers.  Bo Wang’s Galaxy Impact, an ad supported app with over 160,000 downloads, showed an eCPM (effective CPM) of $0.23.  App developer John Kelsey says he sees about $0.50 CPM. Pinch Media CEO Greg Yardley’s “appstore secrets” presentation reports a typical CPM range of 50c – $2 CPM (slide 24) and then in the comments section, Greg quotes developers saying ad rates had dropped to $0.38 CPM. Another developer running CPC ads says he sees $0.01-$0.03c / click.  The truth is that “Most Ad-Funded iPhone Apps Don’t Earn Enough To Buy A Sandwich”

Why the discrepancy?  Fill rate is partially responsible.  Even if a $30 CPM premium ad does exist, it’s not going to run in your app 100% of the time.  In fact, most of the time apps displays remnant (ie NOT premium) ad inventory.  As one developer says “NO ONE can maintain the fill rate at decent cpm”.

Additionally, every ad network wants to attract app developers bad.  Really bad.  So they pitch journalists with juicy stories of high CPMs and ‘case studies’ on developers making sick amounts of money.  Greystripe gets a press piece penned about an “iPhone Beer Pong App Making $7,000 A Month From Ads.”  Adwhirl gets Techcrunch to write “Just How Much Money Can Free iPhone Apps Make? Quite A Bit” which claims apps can make $5000 a day.  And Medialets highlights their $20-30 CPM Dockers ad.  These are the exceptions rather than the rules.  It’s marketing as usual.

The purpose of this post isn’t to point fingers at the ad networks or accuse anyone of lying.  I love ads and I love free apps.  And I love the entrepreneurial spirit in these impressive iPhone ad network companies.  But there’s an important message here for first-time app developers: if you’re considering quitting your cushy job to make $5,000 a day with a fart app, don’t do it.  Always run your back of the envelope calculations first, and don’t assume your app will get anything higher than a $0.50 CPM. Basing your assumptions on $30 CPMs will leave you high and dry.

As usual, readers, I love hearing your comments and questions.  So don’t be shy!

In a sleepy corner of Brooklyn, a technological revolution is taking place.  DUMBO is now home to over a dozen of New York’s hottest startups.  Among them is Sawhorse Media, the company behind Muckrack.com, which took a big step today towards revolutionizing the modern press release.

Early this morning Muckrack began selling “one line press releases” (1lpr anyone?) — twitter-style short form press announcements.  Muckrack.com is a site that aggregates and categorizes the tweets of hundreds of journalists.  Among its loyal visitors are journalists using the site to keep tabs on their colleagues.  By purchasing a 1lpr, you essentially get your message in front of journalists in a form they’ll actually read.

In its current implementation, these are basically sponsored advertisements.  However I see today’s product as part of the first move in a series that will forever change the world of PR.

I first took an interest in PR trying to learn how best to do the press outreach effort for Exit Strategy NYC.  I attended Internet Week’s PR for Startups event and quickly got it:  journalists are drowning with information overload.  As Allen Stern said that night, journalists get pitched hundreds of times each day.  The vast majority of the pitches are misguided and border on spam.  To the startup struggling to get their announcement noticed, relationships are golden.  Second best is identifying the exact journalists who cover your niche and then writing them short, personally tailored emails which are easily scannable.

A few weeks later, I saw a company called MatchPoint present at New York Tech Meetup.  The audience couldn’t have cared less about this product.  But to the people who *got it* and understood the core problem that MatchPoint is attempting to solve – PR professionals struggling to identify the right journalists, and journalists struggling with the information overload caused by mismatched PR pitches — the presentation was revolutionary.  MatchPoint is a communications tool designed for the PR Professional to “help identify and interact with the journalists and bloggers who may actually care about what you have to say.”  Given a press release or several keywords, the software engine produces a list of journalists, ranked by relevance based on a large database of their past news sources.

What I find fascinating and revolutionary about both Muck Rack and Matchpoint is they’re two different solutions to the same signal-to-noise issue plaguing the world of press releases.  Matchpoint solves it with a smarter matching engine.  Muck Rack solves it with bite sized pitches (reinforced by their pricing model: $1 per character with a $50 minimum) — ie constraint and smart pricing.

Right now, both companies are only halfway there in their attempts to solve information overload.  Their products currently give PR professionals a better way to get their message out.  But the real value will come from getting the journalists on board too.  When journalists start trusting these services as reliable sources of personally relevant information, they will become extremely valuable. Once the journalists get on board, there’s a nice network effect and lock-in that will make these services worth millions.

Silicon Alley Insider calls 1lprs the “smartest development in public relations since the canned quote.”  I couldn’t agree more.