Imagine, if you will, a game of Family Feud – but instead of families and random questions, it’s teams of corporate workers and business topics. A grim prospect, I know, but bear with me. If the question, ‘What departments are most people focused?’ came up, I’d bet money the board would be filled with answers like human resources, recruitment, coms, and perhaps marketing somewhere near the bottom. I’d be willing to bet even more money that none of the data disciplines would be up there, backed by flashing lights and heralded by upbeat buzzers. Steve Harvey’s counterpart would likely laugh and throw himself against one of the lecterns, wiping an imagined tear from his eye, if someone were so foolish as to say, ‘Analytics’ or ‘Data Governance’. And that’s understandable; Data professionals have a reputation for being analytical and introverted. I’m inclined to argue that no profession is a monolith, and that some of the most gregarious people I’ve ever met are in data. However, as a data professional who is (with admittedly varying ratios of affection and frustration) often called ‘the resident extrovert’, I will grudgingly acknowledge that the average data professional is perhaps less inclined to leap up and call themselves ‘a people person’ than your average communications specialist.
This perception of data, and indeed data professionals, not being people focused isn’t just something which pervades the social aspects of work environments, it impacts business delivery too. On one side of the coin, we have stakeholders who often view data as this esoteric, terrifying beast to be wrangled only by ‘numbers people’. Attempts to delve into the complexities of data environments or tech solutions are often met by at best polite discomfort, and at worst glazed over eyes and clandestine email checking on separate screens. This reaction isn’t helped by the fact that data teams sometimes face into ‘soft skills’ (I hate that term, but that’s a tangent for another article) or people heavy interactions with similar scepticism. I’ve interacted with data teams from pretty much every possible angle – from customer, to regulator, to team member – and in every instance I’ve encountered some version of the argument ‘I didn’t get into data to deal with people’.
This attitude, that data isn’t a people business, needs to change.
To make it clear (and to avoid losing too many readers), the point of this article is not to Tedtalk all data professionals into extroversion. Instead, I want to drive home the point that, like it or not, data is a people game, and the more you accept it, the more you set your teams up for success.
Why is Data a people business?
To unpack the potentially subjective ‘why’ of the need to shift our mentality, we should first unpack what we know to be true objectively:
1. Numbers in a vacuum are meaningless. It’s the people that interact with the data who give it value.
2. With very few exceptions, regardless of if our job is a data engineer, analyst, developer or governance expert, there is going to be someone who is not a data team member on the end of the transaction chain.
3. Often the hand which pulls the purse strings in a business is not attached to a data person.
So, we know that data is something done for people, by people and is funded by people, specifically in the first and last instance, people who are not from within the cloistered halls of data. This in and of itself is pretty compelling. But beyond that, environmental changes are going to increasingly push us to break down those boundaries, and many of the best and brightest are already making the shift.
Forward-thinking corporations are not only implementing stewardship programs and setting strategic goals around being data led, they are moving towards federated models of data ownership. This doesn’t necessarily mean their sales teams or logistics managers are expected to own technical delivery or development of data lakes. It means all their leaders are accountable for the quality of data when it enters the pipeline, how their teams impact data, and the risks attached to it. We can even see this change reflected in technologies on the forefront of data curation – for example, solutions like Microsoft’s Purview are being delivered with business domain ownership and stewardship practices baked into the naming and structuring conventions.
This move makes sense. Not only is having centralised data ownership incredibly expensive (with data teams required to scale up proportionally to meet business data appetites), but it creates bottlenecks around data teams, encourages complacency when it comes to bad data practices, and disincentivises the development of a mature data culture. Moreover, we see this decentralised model not only succeeding, but so deeply engrained as to be seen as common sense in other functions. The legal team isn’t accountable for not breaking the law for teams; they are accountable for informing them of their compliance obligations, developing assets, and handling the technical component of the legal function. The risk team doesn’t own every risk in the business; they own the tooling and giving the business the insight they need to consistently analyse, own, and mitigate risks within their environment.
To be clear, the issue of data accountability doesn’t always manifest as explicitly mandated ownership by data teams. More often there’s simply not any overt ownership of data as it enters the pipeline at all. Some business leaders haven’t even considered how their team’s actions and processes might impact data. As such, because they are accountable for outcomes at the end of the pipeline and the pipeline itself, it’s assumed data teams have ownership of the whole lifecycle.
This lack of broad accountability for data ownership poses a huge risk for the implementation of technological solutions and advanced analytics tools. Most of us will have heard the phrase ‘data is the new oil’. And indeed, like oil, when the data industry boomed in the late 90s and early 2000s, there was a real inclination to hoard as much as possible – often without consideration for the sustainability of the practice. Now, businesses are feeling the pinch of those behaviours. It’s rare to pass a quarter without a new company – large companies with the scale and financial means to do better, mind you – making headlines because they were the subject to a cyber security incident. Each time the messaging is the same – bad situations made worse because of poor data retention and segmentation practices. This results in stories about customers who leveraged a service 18 years, and two parent companies in the past, shocked that their private information was still there to be taken, or malicious actors given carte blanche access to sensitive data because there wasn’t enough rigour around what should be stored where and if it was truly needed.
I believe we are also beginning to see a trend of regulators leaning on companies to do the right thing – break their poor data habits, stop hoarding information and implement proper controls around how data is retained, manipulated and leveraged. In 2020 the New Zealand Privacy Act was updated for the first time in 27 years, and the Office of the Privacy Commissioner extended their provisions to financially penalise noncompliance, and mandate reporting in instances of serious harm. In March 2024, the Financial Accountability Regime Act 2023 (FAR) was applied to Australian Authorised Deposit-taking Institutions with explicit intent to strengthen accountability among all involved senior leaders for FAR compliance. In the context of the Australian Privacy Principle (APP), APP 11 compels entities to take reasonable steps across the whole spectrum of operations to protect and appropriately curate data, not just data teams. The list goes on and from my perspective the messaging is clear – the regulators already view data accountability in terms of the whole business; we need to catch up.
The ramifications of limited data ownership aren’t just centred around compliance. As I’ve alluded to before, data quality isn’t the sole accountability of data teams, or at least it shouldn’t be. Lack of business processes which uphold data quality are just as damaging to data outputs as poorly designed reports or systems. You can have the most beautifully designed report and savvy data scientists in the world, but if your call centre manager doesn’t assure there’s a process compelling their sales reps to always submit their sales for the day before a report is generated, you’re never going to have an accurate representation of sales on a given date.
This issue only stands to be exacerbated by the implementation of AI (you didn’t think you’d make your way through a data article in this day and age without mention of AI, did you?). I don’t have the word count or the time to fully articulate how broad the benefits and risks of AI are to a business. Further, if this is the type of article you’d read out of interest, you’re likely already awash in the perspectives of both the panicked and the passionate about AI. As such, for our purposes I’ll keep it brief; AI is, among many things, both a magnifier and a concentrator. It increases a team’s ability to compute and broadens access to information, whilst also taking enormous swathes of data and easily distilling them down into small, consumable chunks. This means that, unless you control for it, the implementation of AI tools also magnifies and concentrates existing issues in data. For example, if you don’t have good access control, your teams will be able to access more sensitive information at a greater pace than they would have before. If you have low quality data you will be generating a greater number of inaccurate insights from that data, often without the expert scrutiny of an analyst to flag when something doesn’t feel right, and so on. Because data teams aren’t the only people in a business using AI, they cannot be the only ones positioned to understand these risks and respond to them accordingly. At least not if you want to set your business up for success.
And this is where we come back to people. I know we’ve all heard ‘take them on a journey’ ad nauseam in our places of work over the last few years, but that is absolutely true of the challenges facing data professionals going forward. So many of the best data deliverables fall down and data incidents snowball, not because of lack of technical expertise, but because of culture. We cannot deliver data assets into environments which do not have processes to assure the integrity of the data feeding into them, and expect them to deliver quality data. We cannot build cost effective data functions which won’t be hamstrung into dysfunction with budget cuts at the first economic turn if we don’t decentralise accountability. We will not successfully manage the risks and demands of the next generation of data tools and regulations if we do not broadly enable our people to be good data citizens. People are just as much key enablers for successful and economically viable data functions as tools and technology. In 2022 the World Economic Forum found that 95% of data breaches can be traced to human error. Human error also consistently ranks in the top 10 causes of data quality issues, with the other main offenders such as duplicate data, timeliness issues and inaccuracy all able to be ameliorated if not completely mitigated by robust processes before the data enters the pipeline.
Whether we like it or not, times are changing. From behemoths like Microsoft to small SAS startups, tech solutions are increasingly being delivered with federated data ownership in mind. Human error and non-compliance are some of the key risks to our data security and integrity. Getting business leaders onboard when it comes to accountability for data is at the heart of improving data culture and outcomes. People hold the key to the success or failure for our teams. So, in my opinion, it’s time we all started viewing data as a people business.
About Amy:
When I finished my MSc I had every intention of steadfastly pursuing a career in my field of research- human factors/cognitive engineering. However, partly out of an unexpected love for the field, and partly out of a fully expected love for paying my bills, when my first opportunity as a digital performance analyst popped up, I was hooked. I might end up a data life-er.
Since then, I’ve been lucky enough to work in many areas related to data and tech – Analytics, compliance risk, privacy and a few stints in data governance, including my current position as a Data Governance Lead. Be that as it may, regardless of how technical my positions may get, I still find ways to apply a lens of industrial psychology and human centred design to what I do – technology changes a lot faster than the behaviour of the people powering it.
See Amy’s profile here.