The Power of Context - Data Tipping Point Part 4

How to address data graffiti and fare dodgers and create an environment where data culture flourishes.

This is the final part of a 4 part series on data tipping points. You can start the series here with an introduction to data tipping points - the point where an organisation has enough momentum and leadership to develop a competitive advantage from data without the need for a central data transformation or consultancies.

---

At every data leadership forum, there comes a point where the room concludes that if only the board were more… well on board. If only the board were proactive data sponsors, it would be much easier to transform their organisation. What data leaders are getting to is the power of context, the ability an environment has to shape a movement.

The Power of Context is the third and final of Malcolm Gladwell's rules of tipping points. In Gladwell’s context, this rule explains how small environmental changes can cause a dramatic shift in behaviour - a tipping point.

Those data leaders are both suitable and, at the same time, missing something. Data transformations are so much easier with board engagement. However, data leaders have the opportunity today to provide the leadership they're looking for, earn their board position and transform their organisations.

As we explore two stories on the power of context, I will suggest actions that help create an environment in which data culture thrives. If you've got the remit, take the lead and take these actions. If you don't have the remit yet, then treat these as talking points to start a conversation with your board about the role they need to play in creating a modern organisation that competes on data, analytics and AI.

Broken Windows Theory

The Broken Windows Theory suggests that visible signs of minor crimes, such as vandalism or littering, if left unaddressed, create an environment where more serious criminal activities are more likely to occur.

As a theory, it makes sense. We all know that we're more likely to leave a spotless workplace tidy as we go than we are a messy one.

In 1984, when David Gunn became the New York City Transit Authority president, New York was plagued by violent crime. The subway had an awful reputation; services ran late, graffiti-covered almost every subway car and muggins, and assaults were common.

Image of New York Subway in 1980s

Gunn oversaw a multi-billion programme to rebuild the subway system and attracted criticism for his focus on first removing graffiti, which people equated with arranging deckchairs on the Titanic. However, according to Gunn "The graffiti was symbolic of the collapse of the system… Without winning that battle (against graffiti), all the management reforms and physical changes just weren't going to happen"."

He started a programme to remove graffiti systematically and took the approach that cleaned cars should never be allowed out in public vandalised. They set up a cleaning station at the end of one train line and stopped graffitied cars to clean them during the turnover.

The approach was a success, and in the second phase William Bratton, another disciple of the Broken Windows Theory, took a similar approach to reduce crime on the subway by declaring war on the most minor infraction, fare dodging. Even though a fare only cost $1.25, he put as many as ten policemen at one station to arrest faredodgers and leave them handcuffed until they had a "full catch" to take to the station. Not stopping there, every faredodger was searched (one in twenty carried a weapon) and had a background check ran (one in seven had an outstanding warrant). Unsurprisingly crime on the subway plummeted.

Broken Window Data Theory

So what does this all have to do with data? Data culture is an environment where data is associated with value and managed to deliver value.

This is not the reality in most businesses. Most data work aims to do little more than create reports and is done in poorly managed spreadsheets. Reports by themselves produce no value. But far worse, many of those reports are no longer used, yet the handle continues to be turned.

In this scenario, the faredodger is the stakeholder that asks for a report, or other data project, without articulating the business need or value case. In my experience, these faredodgers exist at all levels, including the board. The graffiti comes in several forms; the spreadsheet whose name ends in "version (2) (4)", the cloud-based data environment made up of poorly named tables with no metadata, the analytics platform with no consistent site structure etc.

It is hard for anyone to stay motivated about data transformation when they work with data faredodgers and are surrounded by data graffiti.

Tackling data faredodgers

As a data transformation leader, I was fortunate to work for a brilliant, value-focused Chief Transformation Officer who challenged everyone to explain the value they were getting from data. At senior levels, if someone proudly showed a dashboard or analysis their team had done, his response was to ask, "but what have you done?".

The point he was getting was that a dashboard alone achieves nothing. To get value from data, you have to do something with the analysis e.g. use a sales analysis to improve stock management and free up working capital.

With his support, data successes that delivered value were celebrated in senior forums, and people soon learned to at least have a plan for realising value (or a request for help to realise value) before showing off data projects as complete.

Tip 1: Explain to your most senior sponsor that they can improve data culture and business outcomes just by setting the expectation that all data work has to be associated with value. All they need to do is celebrate the successful delivery of value from data and press people for their value realisation plans on all data work - from BI projects to data quality programmes.

When I made Chief Data Officer, I continued this mission and told analysts that they had my support to decline data work requests until their stakeholder was clear about the value case for the work.

In hindsight, I needn't have waited to make CDO to take that step. Whatever your role, plenty can be done to tackle data faredodgers.

One simple step we took early in the data transformation was sending out a best practice template for Tableau. There is a little new in doing that. But rather than focus on visual best practices, we also included a cover page with a section on the user story, value case and success criteria for the analysis and dashboard.

Tip 2: Ensure that all best-practice templates for data work focus on the technical and business aspects of best practices. Every product should have a published user story to explain the intended users and use case, including success criteria that detail how success and value will be measured.

Removing data graffiti

Let's first distinguish between data graffiti and imperfections.

Not every data imperfection should be regarded as graffiti. It is just as ridiculous to advocate for cataloguing all data, complete with data dictionaries and owners, as it is to expect every subway car to be sparkling with polish. Just as transportation systems limit their aims to providing safe and effective transport to critical locations, data management should aim to provide effective and trusted access to critical data sets.

Therefore, Data graffiti is evidence of poor data practices that make critical data less accessible and less trusted.

With that aim in mind, the ask of senior stakeholders is easy.

Tip 3: Explain to your stakeholders that they have a critical role to play in creating an organisation that competes on data. All they need to do is ask where the data came for their critical reports came from. Did it come from well-managed and owned systems or unreviewed, unowned workflows and spreadsheets? If the answer is the latter, then ask for this to be solved before asking for any more improvements to the report.

I've seen this work, and one board member drove a data quality campaign in their function by taking this approach.

However, data graffiti is best addressed by the owners of data platforms. Just like New York prevented subway cars from leaving with graffiti on them, data platform owners can put in gateways to prevent poorly managed data from being widely circulated.

Tip 4: Review the structure for your data platforms, from Snowflake to Tableau, and ensure that you have a design with sandpits for development and folders for published work where the folder denotes the owner of the work. Put usage limits on the sandpit (e.g. items in the sandpit can be shared with max five people) that limit work in sandpits and put in a gateway that enables folder owners to control what gets published.

Power of Context Conclusions

The environment in an organisation is likely the single most significant influence on whether data work is value-focused and properly managed.

However, data environments are often so poor and backward that it can feel overwhelming to change the status quo.

The trick to creating a positive data environment is to stop worrying about the number of data faredodgers and the amount of data graffiti and instead start to model the environment as a system. With systems, it is possible to make a dramatic impact by changing key points, like stakeholder reviews and publishing processes.

Tipping Point Conclusions

As we wrap up this series on Data Tipping Points, let’s take a moment to revisit the critical insights we've unearthed together.

We’ve looked at the Law of the Few and the disproportionate role some people play in movements. We explored the Stickiness Factor and how impactful and memorable communication is about more than just facts. Here, we focused on the Power of Context, how the environment can dramatically shape whether or not our efforts will hit that elusive tipping point.

These elements do not operate in isolation. The influencers, the messages, and the context are intertwined, collectively creating an environment where data can either flourish or flounder.

So, what’s next for you as a data professional?

1. Identify Your Influencers: Start with people. Who are the individuals you need to drive change - think beyond data and look for those with networks, deep knowledge and sales skills.

2. Craft Your Message: Remember the importance of ‘sticky’ communication. Ask yourself what you want your audience to think, feel, and do—and shape your message accordingly.

3. Understand and Shape Your Context: Don’t be dismayed by a poor data environment. Understand the data system in your organisation and identify the point where small changes will have a disproportionate impact.

4. Embrace the Journey: Data transformation is a cultural shift, not just a technical one. It’s about changing mindsets, attitudes, and behaviours. Treat it as an epic hike, not a sprint, and take your first step today.

Your role in this isn’t just as an analyst, engineer, or strategist; it’s as a change agent. You have the insights and tools necessary to guide your organization toward a more data-driven future.

Remember, reaching a tipping point is not an end; it's a significant milestone on a longer journey towards becoming a truly data-driven organization. It's about creating a sustained culture where data informs decision-making at every level.

Thank You for Reading

If you’ve got to the end of article four - thank you! This was my first deep dive with a series on a single topic. The next few articles are all planned as standalones, but I’d welcome your opinion on the next deep dive.