Articles   Blogs   News  
 

The Mysterious Case of the Missing Amphibious Vehicles

Or why your data strategy needs to change the lives of the people who care

The following is a tale of amphibious vehicles and data strategy, written by Guy Bradshaw, Consulting Director at Comma Group

I had a strange encounter with a UK police force a couple of years ago. One that revealed a curious ‘fact’ about the vehicles in the county’s primary city. A frustrated officer, who had only recently transferred from front line duties to take a more back office role, complained about the state of the data his Force relied on. He reported that, according to "the system", the city’s most common car type was 'Amphibious Vehicle'.

En route to the meeting, I hadn’t noticed a particular prevalence of Duck Buses or James Bond-style Lotus Esprits avoiding the inevitable traffic jams by launching into the nearest river and, of course, this was never the case in reality.

It was a problem of data.

Bus

Why were amphibious vehicles so popular?

It didn’t take a great deal of research to uncover the root cause: when recording vehicle crime on their handheld device, officers were presented with a drop-down list, neatly sorted in alphabetical order with ‘Amphibious Vehicle’ as the default selection at the top of the list.

Faced with potentially stressful events playing out in front of them, rather than diligently scrolling down the lengthy list of options and casually musing on whether the vehicle in front of them was a ‘luxury saloon’ or an ‘upper medium executive’, officers would simply skip the step focusing instead on more immediately relevant details such as registration, make, model, colour or ‘am I dealing with a dangerous individual?’.

Hence the artificially high rate of ‘Amphibious Vehicles’ being reported.

Is this a data quality failure? Partially, but more fundamentally it is a function of a failed component of a data strategy that should have prevented this choice ever being presented to officers in the first place. Officers agreed that there were times when the ability to record ‘vehicle type’ was highly relevant- when the vehicle registration was missing or obscured for example- but a drop-down list was never going to be a satisfactory way of recording this information. There was also a conviction- almost certainly true- that nowhere in the Force was the ‘vehicle type’ data analysed or reported on meaning that confidence in this particular data item entered a downward spiral. It was hardly used and this fuelled a general viewpoint that the data in the system couldn’t be trusted.

Were this example a one-off instance of a data issue, it may seem a minor, even vaguely comic distraction. Unfortunately for our hard-pressed public servants, this is not an isolated example but a scenario that is played out with regularity.

Another police force mistakenly deployed their drug squad officers to a bust of JCB diggers, a mix-up caused by a mis-interpretation of the offence code ‘Plant Crime’ -‘plant’ decoded as ‘cannabis’ rather than ‘heavy plant machinery’.

The problem of poor data management

The problem of poor data management is not limited to the public sector– I have encountered an international finance conglomerate who eagerly reported three times more unique customers in one country than there were citizens, a global manufacturer with an identity crisis because they didn’t know if their services accounted for more revenue than their manufacturing and a global banking giant who were very proud of their ‘multi-pillar data governance and quality workflow documents’ until ‘last viewed date’ was observed as being over two years ago - hardly a case of putting quality data into the hands of their employees.

Serious data quality issues and, more broadly, poor data management standards, are experienced across organisations in all sectors- large and small, public and private, physical and digital. These problems were not new twenty years ago but today the profile of data in our boardrooms and with the public is higher than ever and the tropes of ‘data is the new oil’, ‘the data-driven business’, ‘data as an asset’ and so on continue to abound. All the grand strategic ambitions declared by organisations- the ‘Big Bets’, the ‘NextGen’s, the ‘OneVision’s etc feature data in some respect and most will claim to put data at the heart of their strategy cycle.

Is it too much of a leap to connect these anecdotal data quality issues with failures of data strategy? No, it isn’t. The ‘Amphibious Vehicle’ example alone exposes a total lack of thought about the value of data and how it is being used. It betrays an absence of connection between critical front-line users and those responsible for designing and implementing the solutions that generate and manage data.

I once found myself invited to review a recently published data strategy with a view to translating it from high level themes into something more tangible. I flicked quickly past the bland pages of ‘vision setting’ in search of something more meaningful, more practical. What initiatives have already been identified? How will the strategy improve the lives of lines of business, of front-line operatives? What changes can be implemented immediately? How will progress be measured? What key decisions are still to be taken, and by when do they need to be taken?

The so-called strategy did not extend beyond a grand vision and was selling its procurer short. Anyone can download a ‘five pillars of data strategy’ template but if the output isn’t going to instruct organisations how to actually go about making changes, how to make an impact and ultimately, deliver improvement and innovation for your organisation, then it is wasting everybody’s time and money.

So, what do I look for in a data strategy?

Data at the heart of your business...and people at the heart of your data.

I am continually surprised at how frequently the people element is not acknowledged as the number one priority for organisations. People are your customers, your employees, your citizens, your business partners, your families… I expect to see the theme of ‘people’ threaded through your business strategy and thus your data strategy. In the Amphibious Vehicles example, who was there to listen to the complaints and, more importantly, the insights about how things could be improved? Some of these changes may be trivial to implement yet could dramatically improve the experience of the front-line user. If the data isn’t working for your people, are you really listening? And if not then why not? You won’t find any better insight into what is working and what isn’t than from those who rely on data to do their job. Such feedback mechanisms must be embedded and instinctive, institutionalised even- and they shouldn’t rely on data professionals alone but be accessible to all.

Commitment from your leaders, engagement from your stakeholders

Top down or bottom up? A successful data strategy needs both executive sponsorship and buy-in from data champions at all levels. Keynote ambitions such as ‘treating data as an asset’ cannot simply mean putting lots of effort into fancy slideware, sitting back and waiting for the magic to happen. It demands serious commitment from executive sponsors and an acknowledgment that, in truth, the journey will never be over. No strategic business plan closes with ‘and in five years the business will be in perfect shape and we can just leave it at that’ but data strategies are often left hanging in the air, a couple of key data initiatives that somehow remain separate from the business, existing alongside day-to-day operations but never really embedded.

The term ‘stakeholder’ is often bandied around freely - I like to think about its literal meaning, ‘those with an interest in the success of the endeavour’. Senior commitment is key but it is the engagement of one or more key stakeholders that really helps to make data efforts tangible- when an interested party from the heart of the business sees the value then you really have found your data champion who will help you drive away from well-intentioned but ultimately empty data maxims and towards a raft of practical initiatives.

To succeed, the data strategy must be fully embedded within the lexicon of the business strategy, the success of which is in turn entwined with a successful data management strategy. Poor data management will hold back all areas of the business. Good data management lays the platform, the foundation, the springboard, for multiple data related innovations.

Undeniably, the alignment of the data strategy to the business strategy is key and yet… for a data strategy to be meaningful it very quickly needs to say something relevant to the business, something that will resonate with those who will benefit most. Which business process will be more efficient? What risk will be mitigated? What insight is being unlocked? It needs to trigger thoughts along the lines of “ah, if this data issue can be fixed, then this process will be quicker, we’ll save time here and we’ll be able to respond to customers quicker there” and so on.

"Don’t make it a data project..."

I was recently asked "what recommendations would you give to an organisation about to embark on a data-related project?" My response was "don’t make it a data project". What I meant was, to be successful, data should be so top-of-mind for business stakeholders that it becomes impossible to undertake a ‘stand-alone data project’. Most organisations will have experienced a failed data governance project at some point in their past, and frequently more than one attempt too. How often have you heard the anguished feedback "we knew we needed data governance but ended up with the data police"?

How are you managing your business without good data?

Measuring actual return on investment is notoriously tricky with data projects but I have a simple rule here: some science is better than no science when it comes to measuring success.

Start with some basic facts and assumptions, build a model of inputs and outputs and lay it open to validation or challenge.

All vertical industry domains feature some form of standards and regulatory reporting and GDPR means we all have data responsibilities now. Thinking about how to demonstrate compliance (or what we have to do) can also help to crystalise thinking about the opportunities (or what we want to do) that can be driven from your organisation’s data. I will sometimes ask customers "what does your data need to look like for you to be compliant?" and the following qualities may be teased out;

"Clear ownership and accountability..."

"Well understood definitions..."

"Good data quality with measures that are tracked and acted upon..."

"Relevant and timely MI, BI, reporting and analytics..."

"Appropriate technology that supports data management..."

My next question is: "but if you haven’t got these key facets of good data management in place, then how are you managing to run your business in the first place?"

Have the confidence to change course

The need for a prioritised timeline is pretty much self-evident. I would highlight again though, the importance of showing early value. A portfolio of quick-win initiatives will grab peoples’ attention early, particularly those who have had doubts about embarking upon the whole data journey. You may also find it useful to categorise initiatives, as Gartner would say, ‘bi-modally’ with, for example, complementary ‘improve’ tracks and ‘explore’ tracks- this will help emphasise the balance that must be struck between delivering early wins whilst simultaneously accommodating long-term progress and innovations.

Internal and external events will buffet your strategy so factor frequent review and flexibility into your planning. Roll back just a few years and you could not move for companies boasting of their ‘2020 Vision’ yet how many of these will have anticipated Brexit, Trump and a global pandemic?

Answer: none, and nor could they in any detail. The point is that the execution of your data strategy will need to accommodate the emergence of both threats and opportunities, and it needs to be able to respond to them both.

So, get the basics right, get priorities agreed, listen to the pains of data consumers, curators, owners and processors, listen to the slew of (sometimes mundane) data problems so engrained that people have stopped complaining about them. Decision-making about short-term vs long-term priorities can be facilitated using a Complexity vs Business Value matrix or similar and I also recommend building a log of key decisions. Key design decisions are (hopefully!) based on sound science with the inputs of appropriate contributors, but months down the line you may find the new Programme Sponsor challenging the choice of technology, implementation approach, supplier selection and so on. With a clear record of why certain decisions were taken (or not taken) you are able to build a comprehensive history of ‘how we arrived here’ and this considered view will enable future planning with confidence- what has worked versus what hasn’t; do other projects need to be delayed or fast-tracked? Should previous decisions be re-visited? Are we still on track and, fundamentally, do we remain relevant to the very stakeholders we set out to help?

A good data strategy will leap from the page enabling you to schedule a complete portfolio of data-related initiatives.

Technology as a feature, not a driver

There’s no doubt that technology has a key role to play in the successful execution of a data strategy. Complex global organisations cannot rely on Excel alone; once, working for a global bank, I was asked to evaluate a data governance ‘roles and responsibilities’ spreadsheet that was literally a matrix of 1000s of lines by 100s of columns. All very worthy, and probably very accurate, but alas pretty useless when it came to informing any behaviour of front-line data operatives. If you’re trying to inspire people to see the value of data and support them in their desire to reach new heights of data capability then to present them with 1000s lines of detail means you’ve already lost that battle.

Whatever your technology options, be aware that there is no ‘silver bullet’, no single piece of software that will execute your data strategy for you. So, if you are facing that ‘what technology shall we use?’ decision point then it still needs to be seen as just one component of that over-arching data roadmap, triggering questions such as ‘what technology do we have in place, can this be refreshed to increase its lifespan?’, ‘what could this tool do for us today, tomorrow, in six months? And if we invest in x, how will that integrate with y in 12 months?’. Having the ability to take the long view means that you can make the right decisions right now, ensuring that investments make an impact in the near-term and also act as the foundation for future evolution of capability and scale.

The eternal campaign - or the battle! - for hearts and minds

You will, I’m sure, at some point have experienced the gap between the hope, expectation and excitement felt internally within a project (and I’m talking any project here, not just data-related programmes) and the deafening silence as it lands with flat indifference in front of an arms-folded audience who have already concluded that this new initiative isn’t going to do anything to make their lives any better.

Counter this from the outset. A coordinated ‘awareness campaign’ can be the decisive factor in transforming a data strategy into one with a transformational imperative. This ‘battle for hearts and minds’ must be in place from day one, not an after-thought once all of the thinking, designing, testing and executing has taken place- and crucially, it needs to continue as your data strategy evolves. Not only will the messaging need to be played and replayed as new stakeholders come and go but the messaging itself will need to continually evolve. The first phase may well focus upon ‘data is important because…’ but as data literacy matures across your organisation, then the messaging will necessarily also have to evolve to a level of greater sophistication. If you can get the messaging just right, you might see that what started as a ‘let’s get a data strategy funded’ campaign morphs into its own thriving community featuring social media, federated governance and powerful data management capabilities that is effectively ‘crowd-sourced’ by its own community.

To be effective your data strategy must be more than a jumble of buzzwords. It must orchestrate a series of sensibly scheduled, high-impact data focused projects that deliver value today but also align to the over-arching strategy. We may have moved on from the ‘2020 Vision’ (replaced no doubt by ‘Accelerate 2025’, ‘Vision XXV’ etc) but the business and data strategies should remain truly symbiotic, in-step and informing each other at all times.

So the next time you are presented with a data strategy, challenge the empty data aphorisms and instead test its ability to deliver a real and tangible impact that will change behaviours on the ground and will make a difference to those who live and breathe the data whether they realise it or not.

Now, where did I park my Amphibious Vehicle?


If you want to discuss your data strategy - or even have a chat about amphibious vehicles - contact us today and we'll be happy to help.