As co-founder and chief software architect for a startup, MyST Technology Partners, I was facing a good kind of problem: client demand for service was outpacing our ability to deliver. We were able to increase our ability to service clients by more than an order of magnitude, without increasing head count. This growth was a result of a process automation framework development.
MyST had developed a general purpose platform technology for building knowledge management applications. As a platform technology, MyST was capable of supporting a wide variety of applications built on the underlying platform. During the early years of the business, we built more than a dozen applications. These applications were either on spec or for specific clients. Two of these gained traction and became the core of our business:
- Enterprise RSS Infrastructure — provided RSS news aggregation and distribution services for large enterprises, including Intel, Verisign, and several telcos.
- Advertorial Marketing Blogsites — provided highly SEO-optimized web sites for SMB clients intended to establish high visibility in organic search. Provides a platform for both inbound and outbound content syndication.
Of these, the advertorial business is the one that presented the scalability challenge. The RSS business, relied on a small number of large clients. Contrarily, the advertorial business relied on a large number of small and medium-sized clients.
We launched blogsite.com to attract such clients and had successfully landed about thirty clients. For each client, we designed, implemented, hosted, and maintained an advertorial marketing website. And, in many cases, we managed the ongoing content creation for those websites. For our small staff (six people), these early clients were maxing out our capacity. Then, one of our early clients, having had great success with their site in the real estate industry, approached us with the idea of launching a spinoff business to aggressively market advertorial services to the entire real estate industry. One way or another, we would have to increase our capacity significantly.
Hiring additional staff and continuing doing what was being done, with more people, is an obvious solution. However, we took a different approach. We asked ourselves how we could accomplish what we were doing with fewer people.
Over a several month period, we developed an abstract advertorial blogsite object model, which we called MyST-X. MyST-X included every aspect of the advertorial solutions we had been delivering. We then developed an XML language schema. This allowed us to describe every client solution in a single XML document, which mapped client requirement into the MyST-X model. Finally, we developed tooling, in the form of a Microsoft word document, with embedded forms and scripting logic, that facilitated the capture of information necessary to produce the XML document for a new project.
At this point, a large portion of the work of bringing on a new client—the task of requirements definition and documentation—had been transformed into a largely self-service process. This allowed us to roughly triple the rate at which we could take on new clients. But, we still had to build and maintain their websites.
Now that the client solution definition was standardized and almost entirely automated, we set about creating tooling that would process such MyST-X object model documents to automatically generate and maintain our client solutions. Having built more than a hundred similar websites manually, we understood a number of design and implementation patterns around which most sites were implemented. We formalized these design patterns and built frameworks of reusable components. Assembling these components programmatically could implement new client sites.
As a result, the time from web site requirements to implemented and deployed solution dropped from 3-5 days to minutes.
Result: >10x Business Scalability
Our approach to scaling the advertorial business involved three fundamental steps:
- Abstract the product definition so that each client solution was a clearly defined instance of a standard solution. There were no custom deliverables, only parameterized instances of standard deliverables.
- Create tooling to largely automate (by making self-service) the collection of client requirements and produce an XML document describing those requirements.
- Create implementation frameworks and tooling. This will largely automate the process of building, deploying, maintaining, and monitoring client sites based on a standardized XML requirements document.
With these mechanisms in place, we were able to launch the RealEstateBlogsites.com joint venture. We grew our client base to nearly 500 clients over the next two years, without increasing our company head count.
Read more about the author F. Andy Seidl : https://www.advantary.co/f-andy-seidl/