Note: This post is a continuation of yesterday’s post.
Follow the money….
While the matters of skill sets and workflow adaptation for a post desktop world are not trivial, it pales in comparison to the real threat that is faced by all classic digital media agencies – the lower revenue-generating ability of mobile application projects.
Here are some numbers I dug up:
- Alex Ahlund, former CEO of AppVee and AndroidApps, and later an advisor to Appolicious, wrote a guest blog article about app sales on TechCrunch. According to that article, a survey of 96 mobile app developers showed the average cost to develop an app was $6,453
- An article on OS X Daily about iPhone Development Costs reported that the development cost range for “small apps” is $3,000 to $8,000 and that “more complex or recognized brand apps” can cost $50,000 to $150,000.
- A well-written article on PadGadget.com explored The Cost of Building an iPad Appand suggested the development costs (as compared to design and other costs) range from $12,000 to $150,000 or more.
- One of the developers of the Twitterrific iOS application has posted an insightful reply on StackOverflow (a programming Q&A site) about the potential development costs of a mobile application. He estimates that a high quality app can run well over $200k during the course of two months, with two developers working 60 hours a week.
Now I know that in a world where you can find any number of companies that will build a very nice web site for only a few thousand dollars, the fact that the pricing of mobile app development is so low should not be a real threat, because professional companies can charge more for higher quality services. But, to use a worn out term, “this time it’s different” – and to explain why I need to go deep into the way-back machine to go back to the genesis of the “digital agency. ”
In the beginning, was the web server, and it was good.
It is hard to overstate just how radical the basic idea of the world wide web was back when it came out. The whole thing started with a really simple communications protocol that didn’t require any pre-arrangement between the client (web browser) and web server – all you needed was an internet connection and the web address of the server, and in a series of exceedingly simple transactions between your web browser and one or more web servers you could gather and place information on a screen with remarkable ease. Add pictures and boom, you’re building web sites.
A web site could make it look like back end systems were actually working together – you could slap an interface together that was pulling information from three different web servers and it would all look OK.

This was utter magic in 1996 – by sticking a web server on the document repository over there, and putting another web server on an image repository over there you could quickly begin to simulate the imaginary seamless “back end system” that the client always wanted.
This was the genesis of what I call “Virtual Technology” – the highly realistic simulation of a well-constructed back office system in which data is effortlessly managed, structured, shared and routed according to enterprise-wide business rules and requirements. The better the simulation, the better the web site.
By the mid 1990s, companies like Think New Ideas, N2K, AGENCY.COM, Modem Media, Razorfish, US Interactive and a slew of others came into their own as large entities able to charge seven figures for a web site not because they could create and operate the back-office systems that their Virtual Technology interfaces simulated – they grew because they could design a way of making the interface act like the middleware between the systems that didn’t really exist. It’s all a lovely simulation because disconnected systems can – in the web browser – appear to be connected. This approach – treating the “back end” as a messy, complex place best ignored and left to others – was (and often still is) the approach taken by many digital media agencies and “web developers.” While there was real progress in making back-office systems directly “web compatible,” back then much of the technology needed to make that happen didn’t exist and was custom-coded.
Then there was a big crash in the digital agency world during the 2001-2004 “nuclear winter” that came after the dot-com blowout, but in this time, new bits of jargon emerged – the “platform” and “Web 2.0″ – of which only “platforms” turned out to matter much.
A “platform” in agency-speak is a way to use a large “back end” system to present and operate a web site that is more operationally optimized and able to connect to other back-office systems “machine to machine.” For example, if you’re in the business of selling widgets, you can pick a platform like ATG or IBM Websphere or eBay’s Magento (or scores of other systems that range from free to millions of dollars), and those platforms will handle things like inventory management, shipping logistics, tax calculation, customer information management, special offer presentation, customer service, offer personalization and much more. They also have “hooks” – little programs – that connect them to things like warehouse management systems or financial reporting systems. Platforms have a repository of user interface templates and content rules that are used to create the web pages that the customer sees. Basically, these platforms are massive databases and programs that access the databases, an it is all presented by putting bits and pieces of information together via templates that become the screens the user sees. Designing a good user experience for a template-driven platform can be a great challenge, but done correctly, it allows for a scalable, manageable online communications infrastructure for any organization of any size.
Great advances in platforms came along, and today, it’s hard to find good large-scale web site that isn’t built using at least one kind of “platform” as a starting point. Even this tiny blog is on a “platform” called WordPress. Platforms make it possible to deploy and integrate interfaces to complex systems with less effort and cost. Hold that thought, it matters much.
So, after a cooling off period after the crash the digital agencies were once again doing very well designing not only the user interface template screens but also the customer-optimized process flows that would be based on the capabilities and limitations of whatever platform the client was using. And still, seven figure budgets prevailed for very large scale web planning and development projects.
That’s because it’s hugely complex to even identify and describe all of the use-cases and interactions that could help a client improve their business performance, much less design the overall user experience or build the templates and make them work.
Digital agencies that got good at helping their clients stay informed about the way emerging digital platforms and tools could be used to change or improve their business or industry were invited by their clients to envision and validate approaches to online communications & commerce, to help the client prioritize these approaches, to structure and visualize the proposed solutions and then to actually execute the projects. The execution of the desktop web project is still, very much, the linear process that I described in an earlier post.
Now two paragraphs ago, I was talking about “platforms” and how they make it easier to build and manage complex systems with less effort and cost. The other thing that platforms do is make it much easier to separate content and function so that any arbitrary device with an internet connection can interface to the platform. In english, that means a web site is only one “version” of the kind of user experience a good platform can support.
In fact, a good platform can support many different user experiences via Application Process Interfaces (API’s) – simple rules that let one computer program talk to another and exchange information. With this model, we reduce the web site to just another interface to the platform, along with many others:

And here’s where it gets interesting.
In the post-PC world, we’re not talking about an interface to a platform as much as we’re talking about one software platform (iOS, Android, Xbox) integrating with another platform (WebSphere, ATG, WordPress) machine-to-machine via native applications (and also having a tight, effective, well crafted interface that suits the device itself).
Here’s the challenge:
Each platform – iOS, Android, XBox, whatever comes out next – has a unique, complex and non-replicating skill set needed to create a great native application (I’ll go into why Native Applications are still better than Mobile Optimized Sites at another time). It’s mostly software development – meaning a well-engineered experience is required, and as anyone who has worked on software projects knows, domain expertise+technical knowledge+dedicated teams=great software.
OK, so you hire people with the software skills you need and drop them into an office and now you’ve got a “mobile practice” right?
Well – maybe not. You see, all of that effort defining process flows and interactions that made these platforms work for the desktop web is very much reusable by the app developer. In fact, they don’t have to – and shouldn’t – use every possible feature and function of the supporting platforms. The post-pc user experience is typically based on a tactical subset of all of the things possible on the desktop web.
And this is where the economics of it all come into play. Smaller shops can be platform specialists – iOS, Android, Xbox – whatever – and they can build great apps that connect to these platforms without the need to define what the platform does holistically or how the company manages the overall platform. That work is left to the digital media agency and systems integrators.
As a result, we see shops that can bid $30,000 for a suite of mobile applications built to operate with the existing web platform while the larger digital media agency can’t even hold a kickoff meeting for that price, much less actually deliver the mobile apps. So the clients give the mobile work to “the mobile shop” or the “xbox shop” – and at the moment, it’s an annoyance. But when you look at the numbers for where consumers are going, it’s mobile, mobile and mobile. And that work is being shopped around much more than the desktop web work.
So, are companies like Razorfish, Organic, RGA and the many smaller shops (like where I work) at serious risk? Maybe. But the digital media agencies – at least the good ones – have something that the small shops don’t, and that’s where I see the biggest change and biggest opportunities.
Part 3: Lights! Camera! Website!