The MA2 story 2: burnt-out, friendless and broke

By Andrew Clifford

If you are a small company, creating a new product is exhausting, lonely and expensive.

Last week I explained why we decided to base the new version of our Metrici Advisor product, MA2, on a radically different design. Our intention was to build a more flexible engine for delivering assessment and analysis solutions.

We have spent a little more than six months developing MA2. Its main features are:

• A secure web-based service, based on the architecture of the original Metrici Advisor, supporting multiple branded offers for multiple users in multiple organisations.
• A general-purpose data structure that can hold any information. This is based on a triple store, with optimisations for performance, security, auditing and version control.
• Metadata (data definitions) held in the same structure as real data, making it easy to fully represent any solution in data, a feature which we call dynamic metadata.
• Automated user interface formatting for viewing and updating the data, which can be manually tailored as required.
• Integration of the jQuery JavaScript library to provide additional user interface capabilities where required. For example, we use it to provide a portal or dashboard interface.
• Addition of server-side scripting, also using JavaScript, to support additional logic. This includes validation, data derivation, user interfaces, web services, and general scripting.
• A charting library allowing data to be displayed graphically.
• Comprehensive security features, such as granular data permissions, HTML filtering to avoid cross-site scripting attacks, exclusive use of prepared SQL to prevent database injection attacks, and so on.

As we built MA2, we began to see the real power of dynamic metadata. We can set up definitions for simple objects, such as web pages, and for complicated objects, such as survey projects, and then set up actual web pages and survey projects, all using the same concepts and the same user interface. Functionality that we would expect to be part of the underlying software, such as account management or product installation, can be represented totally within metadata. We have replicated the functionality of the original Metrici Advisor in a fraction of the time it took us to build the original.

During development, we slowly realised that MA2 is not just an assessment and analysis engine, but is something much more powerful. We have accidentally built a truly general-purpose application.

Technically, I am hugely proud of our achievement with MA2. It is unlike anything I have ever seen, and is fantastically efficient at delivering new IT solutions. However, this achievement has come at a huge cost. We have been working with few breaks for many months. Although our partners, friends and contacts have stayed loyal to us, we have been ignoring them, and have not been seeking out new business partners. Commercially, we have been concentrating on product development rather than bringing in money. We risk being burnt-out, friendless and broke.

The work so far is only the beginning. The next stage is much harder. We have to turn our wonderful new software into a viable product and a viable business. Next week I will describe how we intend to face these new challenges.

Subscription