Planning the Plan
The McFarlan matrix in action
Wherever you are in the world and whatever business you are in, 2009 is going to be a tougher year than usual. If you aren't at the pointy end of the economic downturn then your customers will be. For a lot of IT departments that's going to mean less to spend. For the lucky ones without budget cuts there's going to be a higher expectation of return on that spend.
As it's January, and the start of a new year, I thought an article on planning what to do with those budgets might be both useful and topical. This year, more than ever, I suspect planning (and replanning) skills will be important as we are asked to react to uncomfortably changing circumstances.
Planning can be a very anti-agile activity. I've said before that it's important to remain alert to the fact that there's no inherent value in planning or even in having a plan. The value is all in the execution and delivery of it. Big teams, meetings, paperwork, lots of talking, provide no value if at the end of it there's no time to actually do the work. Good planning should be a light and frequent activity, but the measure of success is all in what you do with the money against what the board of directors wanted.
Richard Nolan and F. Warren McFarlan published an article in the Harvard Business Review in 2005 entitled "Information Technology and the Board of Directors" putting forward a mechanism to more closely align those wants with how the money gets spent. The centrepiece of this work is what's commonly referred to as the McFarlan Matrix (quite why Richard Nolan loses out there I'm not sure, it may just be alliterative misfortune).
They begin with a statement that will ring true to most in corporate IT.
.. most boards remain largely in the dark when it comes to IT spending and strategy.
I don't think this is intentional. I think most boards want to be very aware of how their money is getting spent. I think they want to know that whatever strategy they have in mind is being executed on the ground, in the form of projects that do their bidding. I think the confusion arises because IT strategy, planning and spending is about many things. Not just doing more things, or different things, but also doing the same things better. It's about getting the right mix - at the heart of which is understanding who you are as an organisation.
The McFarlan Matrix
McFarlan and Nolan divided companies into two broad categories: those that require an offensive IT capability and those that may be content with a defensive stance. Companies like airlines, warehousing and logistics operators, high-street retail chains, and the like, may get competitive advantage from having IT systems (try booking an airline ticket using a paper-based system) but they do so by using it to defend their business processes, by making them run more efficiently, more cost-effectively, more reliably and predictably. We appreciate this as customers because we get a slicker service. The board of directors get to drive costs down which pleases the shareholders and us because some of those savings are manifested in reduced prices.
Companies like investment banks, mobile operators, and many e-commerce providers get their competitive edge slightly differently - by using technology offensively to innovate. Of course there's a price for this in that they cannot devote as much attention to optimising business processes, but the markets they operate in tend to punish those that stand still too long with commercial extinction. As customers we appreciate new service structures, new features and new product offerings.
The line is not a hard one - all companies strive for reliable cost-efficient processes, and all want to innovate where appropriate. It's a case of where you spend your budget to best meet the demands of the market you are in. Any one company may, after a few years of innovation, decide to spend a year consolidating market gains and reasserting a reliable base on which to trade.
You can see from the diagram that where you fit is driven by two vectors: need to innovate (new technology) and need for efficiency (reliable technology). Whichever box you are in is where you want to be spending most of your money.
But not all of it.
A clever plan will allocate funds to each box appropriately and will regularly review this against performance and those external factors that affect everyone. A poor plan will allocate funds randomly leaving the board "largely in the dark" and IT staff bemoaning the lack of any direction from the top. In good times the second option may still pay off, but in times like these I wouldn't want to be leaving anything to chance.
Now, I'm not normally a big fan of matrices - so often they are an excuse by the creator to over-simplify or over-categorise real-world situations, which ironically only serves to over-complicate them when it comes to communicating their meaning. I make an exception for the McFarlan Matrix precisely because it doesn't attempt to do this, which makes makes it very useful in planning.
Let's look at a reasonably agile planning cycle and review the usefulness of the matrix.
Our steps would go something like this:
Nolan and McFarlan note in the paper that the percentage of capital budgets assigned to IT varies widely depending on the business model. Over and above 50% in some cases. If you're going to plan properly you cannot start with the answer. What you need in budgetary terms will come from the plan, whether or not it can be afforded is the conversation that follows.
You Are Here
Before you do anything else you have to understand where you are right now, a position neatly reflected by the cells in the McFarlan Matrix.
The cells describe what I would term solution characteristics - that is descriptions of how running systems would affect the business. At a high level this is great because it helps the board understand where they want to be. But we in IT rarely get to design company systems from scratch and so any planning has to take account of today's reality.
A kind of flip-side McFarlan Matrix for this purpose would look like this:
Defensive Offensive Do the same things we do now, but better
Optimise against the median level of service in the market place
Use new technology and approaches to re-factor what we do. Maintain minimal (defensive) risk
Do the things we can't do now. Break new ground. Add novelty
Innovate against the median services in the market place
Use new technology and approaches to revolutionise what we do. Balance risk
Just do it.
These are the things we have to do. Not very sexy but they keep us in business
Use the same technology and approaches to maintain what we do. Seek optimisation only when it's risk free
Fix the broken things
We don't reach the median level of operational efficiency for the market
Use new technology and approaches to transform what we do. Optimisation and efficiency is a bonus but the main focus is on efficacy
Initiatives and ideas will be generated by these four sets of circumstances.
Having categorised all the things we want to do into four broad areas (recognising that this picture is for the current year only, and that some initiatives will deliver against multiple areas), we need to make a first attempt at prioritisation.
Again the McFarlan Matrix helps us with this because we can use it to pitch our ideas appropriately to the higher-ups. Whatever we deliver, it needs to operate at a basic level of effectiveness. This entails good project management, appropriate disaster recovery, financial controls, security, architecture, business continuity planning and so on. Therefore each initiative must be well-formed and thought-through, both to ensure it can be delivered safely and so that it may be understood and supported by the board (or not, if it doesn't fit with the strategic objectives).
What so often goes wrong here is a pile of initiatives are presented, most of which are too tactical to make sense at the high level. There's not enough supporting data for them to appear relevant, which means they are misunderstood and the board opts only for the strategic ones it already understands.
The McFarlan Matrix can be laid on its side to show why this is the case because, as well as representing various operational characteristics and initiative categorisation, it represents the, sometimes blinkered, world-view of the various levels in the organisation. As shown here:
Defensive Support Defensive Factory Offensive Turnaround Offensive Strategic The Workers
Totally Tactical - very practical ideas to maintain things with the risk of making no strategic/commercial progress
Too Defensive - one step removed from on-the-ground activity, seek to optimise with the risk of having opposite effect
Attracted to change - driven to make an impact, desire to improve things with the risk that you lose what already works
Totally Strategic - will to move forward with the risk of being unrealistic and impractical
So to avoid surprises you really need initiatives from one column peer-reviewed by the others. This has a number of benefits - the initiative, if a sensible one, will be more likely to be deliverable at the end of the process (overly ambitious concepts pared down, for example) and similar initiatives may get combined during the review, making better use of resources if they eventually turn into projects.
Understanding these world-views also helps in the presentation of each idea to the board.
For example, if a "totally tactical" idea makes it to the final proposal stage, it's far better to sell the idea along the lines of "in order to be capable of X, we need to do Y", where X is something the board likes the sound of, and Y is the initiative itself.
Making the case that "A, B and C are hopeless, require too many manual workarounds and have shoddy service levels, therefore Y should be funded" both starts with a negative and doesn't factor in the need of the board to feel they are being strategic (to which Y might very well be a contributor).
But we're getting ahead of ourselves, because before we put forward any initiatives there's still two small steps to complete.
We've looked at what we want and where we want to be. That's good, but clever competitors will be doing the same. One way to add a touch of strategic thought to any initiative (or to identify work you haven't yet thought of) is to perform a brief threat assessment. This entails making a few statements about market changes, legal factors, customer needs and what the competition is up to.
You can have the perfect plan to suit your own needs and still have it rendered useless by a sudden shift in the market. Nolan and McFarlan quote the adoption of RFID technology by Wal-Mart as something their competitors did not keep a close eye on. Many of them are now gone and Wal-Mart is the biggest company in the world.
Finally remember this is technology we are talking about. It's a medium that changes by the month. The cost of functionality is dropping all the time. All the skills and approaches and platforms you invested in last year will date very quickly. That doesn't mean they should be changed, but it does mean you have to be aware of the art of the possible.
A few years ago everyone was struggling with data centres and environments. Cloud technology is maturing quickly and for plenty of application areas it's already mature enough to use commercially. At the very least there should be a couple of "in order to be capable of X, we need to run a pilot project around Y technology" in the plan - if you are an offensive mode company I'd suggest this is critical.
The word I have studiously avoided using until now is Portfolio, though what I have described is really just portfolio planning. As in the financial world, planning is a game of risk management. You bet on one thing being worth the investment, based on some careful calculations, and you hedge that with another investment.
That's what the McFarlan Matrix promotes - a couple of punts in the offensive column should be offset by some appropriate defensive activity. The final result is a big list of ideas, concepts and initiatives. Each is a potential capital investment (a project) or something to support via operational business-as-usual development.
Because you know where the board sees the company, and you know where each project belongs on the grid, you can start to do the actual prioritisation. There will always be more to do than you can afford. And even if you can afford it there will be more to do than you can resource, or risk in terms of overall change to manage in one year.
Now it's OK to talk about budget. You can start with the figures the board are comfortable with (usually some variant of last year's spend), apportion percentages to the cells on the matrix, shift initiatives in and out until the result is both manageable and realistic. How long this takes is indeterminate. Ideally no more than a session or two, but alas the final answer is as much about politics as it is about planning. As I said at the start, it's important not to get too hung up on this at the expense of performing useful work.
Planning is a continuous exercise and very soon you will need to be thinking about doing this dance all over again.
Next time we'll talk about sizing projects. It doesn't quite fit here, but it's probably the most important factor in debating whether an initiative should be on your list or not.