Aligning investment with reality


An interesting article appeared at the end of last week, highlighting a shift in thinking towards IT projects within the UK government: was launched quickly and iteratively, with a new simplicity that has resulted in a website containing fewer than 10% of previous separately hosted pages and is set to save as much as £70m on previous arrangements.

A new online system for people to apply for Power of Attorney on behalf of others had taken 10 working days to procure, 24 days to build and code a prototype alpha system for live testing and a beta was due to go live in two months, he said. The whole project had been commissioned from a small business using the G-Cloud for around £50,000 a year, compared with a quote obtained from one large existing provider of £4m set up plus £1.8m per year

There are two reasons why the costs to deliver an IT project should drop significantly to the established normal. 1. At the tail-end of a previous disruptive innovation, 2. At the beginning of a new one.


The hype curve of innovation applies to just about any industry. A new concept is invented. It starts small and is highly specialised but creates demonstrable value to customers. Such success never goes unnoticed and demand begins to grow. That brings competition and rival proprietary solutions. To begin with, more value is created, usually at an increasing pace as different companies come up with more innovative features to compete with one another. But then a tipping point is reached and, for a short while, everything gets a bit chaotic and messy. Hidden costs emerge, problems arise, competitors get acquired and solutions are suddenly discontinued or dramatically altered. Out of the disruption comes a new demand – standardisation that allows for continuity and economies of scale. And so the market settles down into slow growth, cheaper solutions and small incremental improvements. Until a new disruption comes along.

Building traditional web sites for publishing content are at the end of 15-year innovation cycle. The standards for design and formatting of web content have become so well established that even Microsoft has just about embraced them within the latest versions of the Internet Explorer web browser. Most popular public-facing web sites now follow familiar conventions regarding navigation and page layouts. To consolidate multiple different government departmental web sites under a single umbrella web site makes absolute sense and should save a lot of money.

Using agile approaches to software development has grown in popularity in recent years. The goal being to do ‘just enough’ design to build a working solution, and quickly tweak and iterate based on actual usage patterns rather than predicted requirements. It requires far less ‘up-front’ investment due to much shorter planning cycles and usually results in far better user adoption rates. But it doesn’t guarantee a cheaper solution over time. That will depend on the iterations and ongoing development.

Applying an agile approach to business systems is at the early stages of the innovation hype cycle. Some solutions are simply brilliant, but growth in competition means some are not. The disruption and hidden costs are yet to emerge and it’s a little early to be celebrating dramatic savings in annual operating costs. I have already seen one government project that I know is so under-costed, it will take the supplier in question into bankruptcy unless they are able to renegotiate down the line. Yes, the bigger systems integrators were insanely expensive in their quote for what was needed. But insanely cheap is a short-lived improvement.

A comment was made by Tom Loosemore, deputy director of the Government Digital Service (GDS) responsible for the projects quoted above:

“We don’t talk to IT departments other than to ask what legacy can offer”

That’s not a healthy comment and was not well received at the conference where it was made. Today’s IT legacy is just yesterday’s innovations gone stale. I think the GDS could look to the automotive industry for how to better embrace IT as part of doing business. The current approach may be saving a lot of money in the short term (and that’s an understandable driver in the current economic climate) but there are going to be consequences.

Seeking innovative ways to use technology to solve business problems is a good approach. But assuming it is a panacea for all IT projects is not.


Beware Dumb Smart Objectives

The Centre for Possible Studies

It seems that the S.M.A.R.T statement is back in fashion, at least in some circles I participate in. For those who have yet to come across this little gem, it has various definitions but all are along the same lines:

A good objective is Specific, Measurable, Achievable, Results-oriented, and Timely

In other words, if your proposal doesn’t have a specific goal, can’t be measured, might not be achievable, can’t be tied to a specific business benefit and you don’t know how long it will take, don’t do it!

In theory, that’s quite reasonable.

In practice, when applied to systems that people interact with, it can lead to poor decisions and mediocre outcomes – ticks the boxes in terms of specific, measurable and timely but with a huge question mark about the value delivered versus what could have been.

And it gets worse. There’s also a bit of a trend to try and estimate the likely return on investment (RoI) to help prioritise projects competing for limited resources.

Simple RoI calculation

Most people would acknowledge that the above calculation is an unrealistic over-simplification. But that doesn’t seem to stop it from being used.

Systems involving human interaction contain a hugely unpredictable element: the organic material found between the chair and keyboard (or floor and tablet). Those interactions impact both the cost to implement systems and the value created (or not), in ways that can never be fully predicted in advance.

Even if humans aren’t involved, cost and value often have non-monetary elements. How do you calculate them? Scores out of 10? Good luck with that. When the government introduces new legislation that your company has to comply with, the value and RoI do not matter. It will override all current projects if a legal liability and deadline are included.

Using S.M.A.R.T to sanity check a business case can be a useful exercise. It makes you focus on knowing why the project is worth doing compared to alternative activities that would take place instead. All projects face some level of constraint be it time, money or resources. Just relax a little on the approach and don’t be too quick to dismiss proposals that don’t comfortably fit a ‘smart’ objective:

  • Is there a specific reason for the project? If not, it’s a vision. Doesn’t mean it shouldn’t be done (or rather, attempted) but be realistic about the knowns versus unknowns involved – executive sponsorship will be highly influential
  • Does the project have measurable outcomes? If yes, great! Just don’t try and estimate the end result. Instead, have robust data about the base line – the current position that needs improving – and the underlying causes
  • How does this project compare to other options in the time and resources required: 1. Doing nothing; 2. Doing something else; 3. Doing this project differently? This is where resource constraints may force some difficult decisions and political factors man skew priorities

Here are two case studies involving systems with a high element of human interaction. One ‘smart’ and one not. Both delivered positive outcomes. Neither needed to calculate an estimated RoI in advance or could have predicted what those outcomes would lead to.

Example 1

In 2009, there were 850,000 recorded incidents in the UK documenting poor clinical handover of patient care that resulted in, directly or indirectly, to 3,500 deaths

Doctors in general medicine ranked Barking, Havering and Redbridge University Hospitals NHS Trust (BHRUT) 31st out of 34 London hospitals

That’s a target to improve!

BHRUT identified that many issues with the clinical handover process were occurring on Fridays and weekend mortality rates were significantly worse than during the week. They targeted filling in the communications gaps between the systems: teams working on different shifts, with different IT systems, and at different locations.

They decided to leverage existing systems rather than introduce new ones. SharePoint 2010 was already deployed, but only as a basic document management system. It became the central portal connecting the different line-of-business (LoB) systems, making it easier to submit updates and providing a single consolidated view of patient care history.

Early results achieved:

  • 98% completion rate of tasks using the new approach
  • Speed of the ‘end of shift’ handover meeting reduced from 45 minutes to 20 minutes
  • Weekend mortality rates began to fall

Just 12 months after being placed third from bottom, doctors ranked BHRUT 1st out of 34 hospitals across London

That outcome could never have been accurately estimated in advance. The project does satisfy the basics of the SMART statement. It had a specific remit, a design that was achievable, outcomes that were measurable and results-oriented, and a project delivered in a timely fashion. But above all else, it had a clear target to focus on.

Source: Microsoft case study, October 2012

Example 2

“At companies, especially technology companies, the most brilliant insights tend to come from people other than senior management. So we created a marketplace to harvest collective genius.”

The above quote is from the CEO of Rite Solutions. They decided to implement an internal prediction market where anyone can submit a proposal that is converted into stocks. Employees can buy and sell the stocks as a vote of approval or not. Changes in the prices reflect the sentiments of the organisation, regardless of role or seniority.

Every employee gets $10,000 in “opinion money” to allocate among the offerings, and employees signal their enthusiasm by investing in a stock and, better yet, volunteering to work on the project. Volunteers share in the proceeds, in the form of real money, if the stock becomes a product or delivers savings.

This example flies in the face of SMART twice. The idea doesn’t meet the criteria. It’s specific – let’s create a prediction market and see what happens. But that’s about as far as the business case goes. Really it’s a vision, seeking out new innovative ways to grow the business.  And the system is all about prioritising projects without any consideration for calculating potential RoI before starting on them.

And it worked.

One of the most popular early stocks was a proposal based on video gaming. It was not popular with senior management who weren’t interested in playing games, “I’m not a joystick jockey.” But support among employees was overwhelming. The product was launched and became responsible for 30% of annual sales!

“Would this have happened if it were just up to the guys at the top?” Mr. Marino asked. “Absolutely not. But we could not ignore the fact that so many people were rallying around the idea. This system removes the terrible burden of us always having to be right.”

What a refreshing approach to management, and not a ‘smart’ statement in sight.

Source: New York Times article, March 2006

Image: I took that photo walking through London last Summer – Possible studies? Answers on a postcard, please

What is the CIO’s mandate?

The following presentation is a condensed version modified for display on Slideshare. It is based on a workshop I run for organisations to help define a digital strategy, by understanding and aligning with current business priorities. I’m considering expanding it out into a book format to include templates and exercises from the workshop.  The content is based on a range of materials I’ve gathered over the years from studies into achieving business value from IT. This version has been recently updated to include statistics from IBM’s 2011 Global CIO Study.

One of the most hotly debated exercises is to take the business benefits diagram and apply it to an industry to identify the differences between market leaders, those struggling to survive and a sample in between. Or alternatively, plot how a company has changed its position over time.

 An easy example is Apple. At the start of the century, the company was not in the greatest shape. There was little focus on customers, products were expensive to buy partly because they were expensive to produce, and the quality had deteriorated from being an early market leader. In the following 12 years, they moved all three axis into dominant positions

  • The opening of the Apple retail stores, staffed by ‘geniuses’ changed the customer relationship for consumers. iTunes provides a treasure trove of data about purchasing decisions. One are that remains weak is the customer relationship in the enterprise space.
  • Operations were completely transformed under Tim Cook, enabling products to be manufactured far more efficiently and at lower costs.  Whilst the Macbook range is still more expensive than the average personal computer, the gap has reduced significantly when comparing like-for-like specifications. And Apple makes much higher profit margins compared competitors, making it harder for others to compete on price.
  • Product quality has returned to a strong position, with most other manufacturers seeking to imitate many of the design concepts introduced in Apple products over the past 6 years since the launch of the iPhone
  • …and the fourth axis – redefining the market. In barely 6 years, Apple’s revenues and profits from mobile devices surpassed their computer sales and turned them into the most valuable company in the world. Nobody expected the oil companies to be toppled just yet.

If you were a competitor to Apple, which axis do you focus on to improve performance?  The fourth is always still there – Apple has yet to dominate the enterprise market outside of a focus on education…