Building a Developer Community

Now is the age of developer programs or “developer communities” as they like to style themselves. Five years ago there was barely a handful of such programs in existence. Microsoft led the way for platform vendors, and if you add Java, IBM, Adobe, and Apple to the mix you’re close to the complete list. Now there are literally hundreds of developer programs; from Paypal to Vodafone hundreds of technology companies offer a developer program.

All these developer programs are created for the same reason:

• More developers –> more applications for the platform
• More applications –> more consumer adoption

This is the formula that gave MS Windows such success in the mid-90s and Apple’s iPhone in the past couple of years.

But as all these companies scramble to build a developer community I see a general misunderstanding of the key success factors; leading to most companies focusing their efforts ineffectively. So I wanted to write-up my view of the key success factors for building a “developer program”. Note that I’m talking here about developer programs offered by platform vendors to the developers creating software around their platforms. In particular I’m not talking about open source projects and their contributor developer communities which is a different kettle of fish entirely.

.

Developer Community

Before describing the key success factors, it’s useful to demystify the term “developer community” as it applies to software platforms:

1. Primarily it is simply the set of developers creating software for the platform.
2. A more genuine ‘community’ exists when these developers support and inspire each other to great better software.

Regardless of what you think “developer community” should mean, 99% of the time this term is used by software platforms, it means (1). They like to imply it means (2), but it doesn’t. Very few platforms have genuine communities.

.

Key Factors

I believe that the key factors for building a successful developer program (in order) are:

(1) Awesome platform. You need a platform that lets developers do things that they can’t do anywhere else. The first platform that enables developers to do really effective Augmented Reality applications will benefit greatly from this.

(2) Channel to users. A simple, low barrier, channel to a large number of users.

(3) Support. The support offering has now been almost standardised with SDK, IDE, example code, reference library, guides, FAQs, and support forums being the main elements.

(4) Promotion. Tell developers about your technology, what they can do with it, and how to access it.

(5) Collaboration infrastructure. Provide infrastructure to encourage and help developers to interact on-line and off-line; discussion forums, events, ideagoras, etc.

Note that while list is primarily focused on application developers; it essentially holds for OEM-style developer programs too.

.

Comments

MS showed us how to do developer support properly, Apple showed us how to provide a channel to users properly (at least in the mobile world), and the social networking craze has given some reasonably good hints on collaboration infrastructure. So most companies seem to get these bits (or at least get that they don’t get them but really should).

But very few companies seem to understand that everything else is meaningless and will come to nothing if there isn’t an awesome platform underneath everything. The channel won’t succeed if there isn’t great unique content. Promotion will become despised if it’s empty. Collaboration simply won’t happen if there’s nothing interesting to talk about.

Why does this happen? Typically companies create a “developer marketing” team to create a “developer community”. This team almost never has any real influence on product requirements. They are a downstream team that take what they are given and do what they can. If they are mainly technical guys then they will focus on providing good support and maybe some on-line collaboration. If they are mainly marketing guys then they’ll focus on channel to users, promotion, and trying to build a social networking website.

Whilst a little subtle, there is a pervasive difference in how you work when creating a product vs creating a platform. This makes it a difficult change and like all change it is generally done badly.

Sony Ericsson and Motorola

Sony Ericsson and Motorola have both been in the  “struggling” camp for some time. From my posting below on Palm’s cash flow, I’ve received a few questions about these two companies and wanted to post up some quick thoughts.

.
Data

The table below shows data extracted from SEMC’s Q2 results and Motorola’s Q2 10-K form.

.

Sony Ericsson

• With respect to my previous post, the first thing to notice is that SEMC and Moto (ailing as they are) are still in a different league than Palm. Palm had $68M of sales in Q2 while SEMC had $2.4B (based on today’s exchange rate).
• SEMC have a lot of cash, more assets, and healthy equity. So basically – nothing is forcing them out of the market. And given that they’ve just appointed a new CEO, I assume that the parent and other investors will give him a chance.
• The most striking thing for me about SEMC’s results is the jump from €200M gross profit down to -€213M net profit. Any CEO is going to see that and look for where the operating costs are being wasted. That’s going to bring SEMC’s broad platform strategy deeply into question (Symbian, OSE, WiMo, and Android).
• My guess is that after the usual 3 months the new CEO will announce a strategy that reduces the number of platforms by at least one, maybe two, and cuts the rather excessive traditional SEMC marketing budget. If SEMC traded separately this would push the share price up for a few months. Then success depends on execution of the strategy and actually producing some good phones.

.

Motorola

• Motorola made a profit! I almost fell off my chair. In Q1 they lost about $180M.  In H1 2008 they lost $190M. Basically I can’t remember talking about Motorola making a profit. Obviously $26M is hardly a massive profit when you have $25B in assets, but it’s very positive news for Moto.
• Unfortunately they haven’t reported Q2 cash flow, so I can’t see if there are any games going on here. They report an H1 2009 cash flow of -$183M. When you consider that they made a net loss of $180M in Q1 then it’s looking like the numbers might indeed be adding up.
• I’d say that the new SEMC CEO will be looking closely at what Sanjay’s done at Moto and might just copy a bit.

.

Palm’s Cash Flow

I’ve never quite understood how Palm’s business plan adds up. Fortunately they released their latest quarterly results (June – August) with the SEC just a few days ago (September 17th) so I decided to seek enlightenment. I didn’t find it. But I did note a few interesting about cash flow that I thought I’d write up.

• Palm made a net loss of $164M in the quarter. I would like to mention that I was surprised at how little this was reported in the industry news. When you compare it to the extremely negative reporting of Nokia’s results you can’t help but feel that the silicon-valley-club conspiracy theories have some justification.
• Palm’s actual net cash flow was -$42M; i.e. although on paper Palm lost $164M, their bank account only shows a loss of $42M. This is primarily due to Palm deferring a majority of Pre revenue and costs ($114M net).
• Palm have $110M in cash and another $100M in short term investments.
• So at the current burn-rate that gives them 5 quarters.
• Palm currently have a current asset to liability ratio of 0.56. They have a total asset to liability ratio of 0.5. They are running a shareholder equity deficit of $740M; i.e. if you sold all Palm’s assets you’d still need another $740M to cover its debts. And if you consider that $166M of Palm’s assets is in “Goodwill” which is quite difficult to turn into cash it makes the picture look a little more grim. So I can’t see too many people lining up to offer Palm a loan or sensible investors buying up a new rights issue. The key point is – I don’t see where Palm is going to get more cash from.

I can’t see big things happening for Palm in Sept – Nov. Then there’s Pixi and Christmas. February to April is quiet. Another potential release in May for the summer market. Then a slow ride into November 2010.

So as far as I can tell Palm’s current business plan is only viable if it turns a good profit over Christmas with Pixie (let’s say $100M retained earnings net profit) and then backs it up again in May. Any deviation from that and I don’t get how Palm can keep going and what fool’s money is keeping it going. A failure over Christmas will inevitably lead to the usual uninformed conjecture about “Dell should buy Palm”, “HP should buy Palm”, etc; but Palm’s liabilities could easily run into $1B. And what do you get for that? Not enough. I’m sure that everything will play out in complete contradiction of everything I think and have said here, and that’s mainly why I’ve written it, so that this time next year I can read this posting and admire the wonderful logical-defying world that is the technology industry.

=======================

22/09/09. I thought I’d update this posting with Palm’s share price over the next year to see how things develop. At the time of writing (22/09/09) the share price is $15.95 (up 14% today).

24/09/09. It looks like I’ll be tested on “I can’t see too many people lining up to … buy up a new rights issue”. Palm are preparing a new rights issue. 20M shares at $16.25 each; i.e. $325M. It’ll be interesting to see what happens.

24/09/09. Palm’s shares hit a long-time high of $18.10 today and then fell back to $16.94. Apparently this is based on a rumour that Nokia will buy Palm. It’s not out of the question (stranger things have happened), but it seems like a bad deal to me. Spending $4B to acquire $1B of debt, an operating system you don’t want and is too hard to merge, and a difficult brand situation. Everyone talks about “US presence” which is certainly a problem for Nokia; but I think they could do a lot better for $4B. Asa  random thought – for that they could give away 20M phones for free! That would make for a nice deal with Verizon, AT&T, or Sprint and get Nokia devices into US hands. Anyway – lots going on with Palm.

28/10/09. Palm’s share price is at $12.59 today. They’ve been steadily decreasing over the last few weeks. That Pixi launch better be very very good or my prediction runs a serious risk of becoming correct and I will likely become intolerably smug.

 

Social Networks

No topic seems as shrouded in misconception, vapidity, and hyperbole as “Social Networking”. Whilst a certain latitude must be allowed given the rapid change in the online social networking world, there is one misconception that seems rather fundamental, is rather pervasive, and essentially annoys me sufficiently that I decided to rant write about it.

Facebook, Twitter, Orkut, MySpace, Bebo, etc — are not social networks.

‘Social networks’ are networks of people communicating. A talks to B who talks to C. When you describe a social network you talk about the individuals involved and what they talk about (e.g. Wii games, modeling of neurologlial communication, mobile computing, or AJAX programming). You probably would include how the network communicates, but that’s more a practical matter rather than a defining feature of the network.

If Facebook and Twitter are not social networks, what are they? They are tools to support social networks. Fantastic tools. Tools that have led to an enormous growth in the number of networks, the number of individuals involved, and the volume of communication within those networks. They are an essential part of the social networking landscape, but they themselves are not networks. They help individuals manage their networks, they facilitate simple broadcast multi-media (in the literal sense) communication, and they typically provide tools to expand network membership.

Why does this distinction matter? Well… for people who simply like to use these tools to engage in different social networks, then it doesn’t.

But for people analysing social networks from an industry perspective, planning marketing campaigns using social networks, and especially those writing articles about social networks, it would seem rather important. A few examples:

• This distinction is what underlies the success of “social network aggregation” software. The fact that users have networks of people and communicate with them using several different tools makes this aggregation attractive. If Facebook and Twitter were actual distinct social networks then aggregation would be of marginal benefit if any.
• This distinction is the key factor behind the current obsession with “owning” the address book. Most major operators and mobile platform vendors are spending serious money in an attempt to “own” your address book since that is really what defines your network; everything else is simply a communication channel. Note that I think much of the current thinking around address book ownership is flawed as everyone thinks about why it would be great for their company rather than how users will approach the problem, but that’s a longer discussion for another day.
• A marketing campaign that targets “being on Facebook” is half-baked. It needs to go further and look for the relevant social networks on Facebook and get involved in those conversations. A marketing campaign needs to understand its audience, know what the active networks are, and then join those networks.
• If you are providing tools to a social network, rather than being a social network, it entirely changes how you think about your own service. Looking at the major social networking website it becomes clear quite quickly who understands this and who doesn’t.

How much is a tooth worth?

Yesterday my eldest nephew lost his first tooth and placed my sister in a quandary —  how much should the tooth fairy pay for the tooth? For those of you from continental Europe, the tooth fairly is the equivalent of the tooth mouse for people who dislike the notion of rodents stealing into their children’s bedrooms late at night (even if they do leave money). Regardless of the species of the purchaser I felt sure that economic theory would provide a conclusive solution to my sister’s predicament.

1) Labor Theory of Value
The labor theory of value (LTV) asserts that the value of a commodity is related to the labour needed to produce the commodity and the costs of any materials (where the cost is itself based on the LTV). So using the information below we can assertain the correct value of the tooth:

• Teeth are made primarily of hydroxyapatite. All other materials are in sufficiently small volumes to be ignored.
• The average baby tooth is 0.5 grams.
• The current market rate for hydroxyapatite (in small quantities) is about 12c a gram.
• Now my nephew is 7, but it’s difficult to claim that he has spent 100% of his energy on growing this tooth. I think a generous estimate is to suggest that he’s spent cumulatively 1 day working on this tooth during his life.
• For some reason I’ve had difficulty finding average salary figures for 7 years old in the western world, but fortunately there seem to be no such problems in much of southern Asia. From these I understand that $3 is a fairly reasonable daily rate.

So based on this I would value the tooth at $3.06 (USD).
.
(2) Marginal Utility Theory of Value
Of course, although the great Adam Smith was a proponent, the LTV is clearly complete nonsense. Interestingly it is also a key foundation of Marxist economics. The western world today mostly uses the marginal utility theory of value which roughly states that a commodity is worth the benefit it would provide to someone. “Marginal” means that it’s not the value of free roaming teeth in general, it’s the individual value that this increase in supply of 1 tooth will bring to a consumer.

Clearly the value on this basis is $0.
.
(3) Market Value
But hold that thought! This is the modern world of financial markets, where commodities are not sold based on their actual marginal utility, but according to the most erroneously high perception of marginal value. Now a quick survey of parents I know shows that 0% of them know how many teeth a child has. So it is entirely plausible that a child could fake the losing of a tooth. Clearly this lays the foundation for a free market in the trade of “lost” teeth.

A brief survey of the Internet shows that the highest price paid for teeth (by enough people to support an active market) is $5 (USD). So the market value of the tooth is $5.

Conclusion
So once again, what seemed like a rather random decision, is actually a fundamental question of your economic beliefs. Furthermore, if you decide that theory (2) is the only conscionable choice, and therefore a value of $0, one is faced with the daunting philosophical question of whether the very notion of using economics to deal with 7 year olds is valid. I leave this as an exercise to the parent.