Agile Business Conference 2009
17 October 2009
I've just returned from this year's conference; some of the highlights are listed below.
Nik Silver of the Guardian, described his organisation's journey to fulfil their mission statement of being the "to become the world’s leading liberal voice." His observations were that:
because their on-line readership was global implied that their website and supporting services had to be genuinely 24/7 operations; and
in revamping their website, the team experienced a 40% requirements' churn.
When asked to examine its backlog, the team found "projects" ranging from "add a new button" [on a single screen] to "help expand in India"!
In addition, the organisation was the first media organisation to supply "full-content" RSS, i.e. newsfeeds with complete content material. As well as intangible benefits such as positive media coverage, tangible savings were made when the development costs of its mobile site dropped to zero. This was because the mobile development team had immediate access to the complete content and didn't have to keep asking the core team to supply additional information.
Next, Iain McKenna and colleagues talked about a recent project turnaround that succeeded only on the third attempt. Some of their findings were:
Changing the reference to "project" to "product" to make the team realise that support was required after delivery.
They had to strictly prioritise high-level initiatives; for example, there was no chance of executing an offshore strategy at the same time as implementing the project turnaround.
It proved essential to re-organise siloed departments into value-streams, resulting in a shortening of the value-chain, for example, a product manager would write "acceptance tests" for marketing ideas.
It proved essential to educate the business to understand that the technical department was a partner rather than just a third-party supplier.
It was important to work with other business areas to ensure all departments were being aligned to corporate objectives. This enabled the team to get warning what requirements were in the business pipeline so that there were no surprises.
As team of twenty-five offshore developers were replaced with a team of four onshore developers.....who, as a team, were 400 % more effective.
The result was that after six months there was a 45% increase in revenues from the product.
Next, Dot Tudor gave an interactive presentation on the use of ITIL in agile teams. Her key message was to involve a stakeholder for each part of the ITIL cycle (eg Service Design, Service Operations) within the agile delivery team...the challenge was to keep the team small enough to work effectively.
Then, Simon Buck of Symbian, and his colleagues described how Symbian (later taken-over by Nokia) introduced agile techniques throughout the organisation.
He gave some idea of the challenges faced; namely that 1000 developers were working on 50 million lines of code with a fixed technical architecture.
The company introduced three process levels:
at the lowest level, a two-week iteration for incremental delivery;
an eight- to ten-week planning cycle to coordinate multiple projects; and
a programme-level planning activity.
The transformation team found that the biggest resistance was from project managers, until the team explained that project managers were essential in the overall process to coordinate the development of coarse-grained features across multiple teams.
Next-up was Mark Simmons who talked about agility in government. He mentioned that OGC processes should be tailored around your own process, rather than the other way round. He said that the key to flexible government contracts was a set of scoping statements which describes what's in scope but also, importantly, what's out of scope. The best estimate of cost at this stage should be plus or minus 20%. In addition, there should be named decision-makers and rapid turn-around times.
On the second day, Stephen Carver of Cranfield University gave an energetic and enjoyable presentation called "Fear of Flying" with compared project managers to pilots and project delivery to air-traffic control.
Some highlights: The comment that "If you living in the present you are not a project manager."
On the RAF's plotting table – the tacticians' view was that of the enemy's. In addition, the wooden markers used, in conjunction with lighted wall displays showing altitudes of aircraft, were an early example of "information radiators."
He equated an air-traffic controller to a programme manager – ensuring that resources are fully utilised. He found that the end of a programme isn't always clearly defined - "we've more or less achieved our objectives, so we may as well close it".
Then Mike Griffiths gave a talk on Earned Value Analysis, as applied to agile projects, explaining that EVA tracks a combination of time and cost elements simultaneously.
He also mentioned that:
You should only track what you want to influence to avoid the problem of inadvertently setting an inappropriate target.
A frequent issue is how to allocate a project's proposed savings to the list of features that the project is delivering. In the absence of any other guidance, one approach is to split savings equally across all requirements in the feature list.
It's not worth performing advanced maths or using sophisticated techniques on fegures that have been derived from little more than guesses; tailor the formality of the analysis to the quality of the data.