Why you need maps and not just routes in IT system design

In their book How Big Things Get Done, authors Bent Flyvbjerg and Dan Gardner highlight the hidden dangers to large projects of using ‘sunk costs’ as a justification to continue. As with a lot of management speak, it’s a modern update to an age-old maxim, in this case, ‘Don’t throw good money after bad.’ In their defence though, they use it to show how the temptation to continue can be especially damaging on very large projects.

A timely case study on this issue is the beleaguered High Speed 2 project. The UK government confirmed this week they will complete the current scope of the HS2 programme from London Euston to the West Midlands, despite the “very significant” increases in cost and delays to delivery.

The ‘sunk costs’ ethos

‘Sunk costs’ and ‘good money’ both imply that just because resources have already been spent on something, it shouldn’t be assumed further spending is necessarily a good thing.

In both cases the monetary aspects of the problem are the focus. But while those are certainly important, the principle could equally be applied to other areas, like ideas, patterns and plans in IT system design.

We all get attached to things we’ve invested considerable time and effort in. And, in turn, that time and effort almost feel like an argument for further effort. So, having developed a concept, made a design or decided what to do, this historical activity makes it all too easy for you to carry on. And that’s even when circumstances, and more importantly results, suggest things aren’t going quite as you intended.

The problem with the stop-go approach

The answer might sound obvious. Surely it just requires that when further resources are needed, you carry out a review of progress so far and then use your discipline and resolve to make sure any later change decisions are seen through. For example, you should stop if you’re not where you expected.

But that black and white stop-go approach has its own problems. Previous effort and time may not be enough to justify continuing. But, conversely, deciding all work so far should be abandoned could mean you ignore the value that’s been created.

Financial decisions may indeed be as simple as, ‘Can we afford to provide more funding?’ However, other aspects are more complex, such as:

• What’s good about the idea so far?
• Do some aspects of this design help solve the problem?
• Where has this plan succeeded?

These could all imply not continuing as is, but equally not stopping entirely either.

Don’t throw the baby out with the bathwater

Another old saying comes to mind, ‘Don’t throw the baby out with the bathwater.’ So, if a point is reached where no further funding is available, you may not even get the opportunity to use what’s been created so far. It’s therefore important to review progress before that funding ‘cliff-edge’ is reached. Regular reviews with multiple alternatives are needed, with ‘stop’ very much reserved as a last resort.

This also sounds easy enough. But, in practice, it’s quite a change.

It suggests that rather than simply understanding a single approach, many other options must be understood too. So, what’s needed isn’t just a route, but a map.

Maps – not just routes

Continuing that geographic analogy, you could argue that when planning a route, a map is a pretty good starting point. However, once you’ve decided on a course and started following it (‘sunk miles’), your original map is likely to be forgotten. At that point it’s all too easy to concentrate on just navigating the chosen path rather than considering the wider surroundings.

Equally, routes tend to be, or at least become, a series of detailed steps, and this can add to the perception your original map is redundant. So, keeping that overall map visible and up to date is important.

But that requires balance. It’s tedious and unnecessary to provide a map at the level of intricate detail you require to navigate a full route. What you need is just enough detail to allow possible alternative routes to be visible, but without so much detail it becomes unfeasible to create (too costly), impossible to read (too busy/detailed) and too localised. Think of the difference in detail between an Ordnance Survey Explorer walking map and a road atlas.

A similar problem exists when describing organisational processes. Operational activities need a level of detail that isn’t necessary or useful when planning, reporting or even communicating between activities. As such, an activity-level map of the organisational processes will be too detailed.

The balance is to use an approach that focuses on the points where the activities are connected: their ‘intersections’ and the routes into and out from them. This means defining what goes into each activity, what’s involved and what comes out of it, but without locking the activities into a fixed sequence. Over time, activities will naturally form sequences, but they don’t have to be fixed.

Understanding system behaviour

Like a road atlas, this ‘intersection map’ gives you a good starting point when trying to understand how a system actually behaves. And when the intersections are clear, not only will the data and detailed logic needed to support them follow naturally, but this view makes it easier to see alternative routes.

The data and logic involved are defined as part of the intersection itself rather than the routes between them. So, a system can grow more organically over time instead of being built around assumptions that quickly become forgotten and increasingly dependent on each other.

Working this way also means detail can be added gradually as understanding develops, rather than having to define everything in advance. It’s a model that supports continuous improvement and adaptation by design.

How NumberCloud and NumberWorks can help

Quantiv’s NumberCloud service and NumberWorks method are both built around this approach. NumberWorks allows patterns to emerge over time and be described in a consistent way, while NumberCloud supports the collection of metrics without needing to predefine what data will be gathered.

Taken together, the goal is perhaps best captured by a hybrid saying, ‘Don’t throw good money out with the bathwater.’ Or, ‘Don’t throw the baby out with the bad money.’

To learn more, contact the Quantiv team on 0161 927 4000 or email info@quantiv.com