Data trends that shaped 2023

2023 was a busy one in the world of data, with the rise of AI, new concepts like Data Products, and challenges to the way Data & Analytics teams prove their value.

2023 has been a huge year for data management. “Clean data” has emerged a significant differentiator as consumers of data-enabled and AI services no longer settle for wonky outputs, and have wised up to the marketers trumpeting perfect use cases with pre-recorded videos.

From the ether straight into the mainstream, ChatGPT (and others like Claude and Bard) are now generating sophisticated outputs from multiple inputs. No longer solely the domain of experts, your average person can now harness Generative AI to produce images and text. Will Prompt Engineer displace Data Scientist as the sexiest job title of 2024?

Though easily the most prominent event of the last year, AI was not the sole focus. We’ve seen teams reshaped, buzz words take flight (hello, Data Products) and businesses realise that data is indeed a competitive advantage. The gap between the frantically paddling swan and the gold egg laying goose is far – take stock of your team’s readiness, maturity and capabilities is the message.

Will Prompt Engineer displace Data Scientist as the sexiest job title of 2024?

AI - the hero of 2023?

Artificial intelligence (AI) has been around since the 1960s in the form of algorithms, and the birth of the Perceptron was even earlier in 1958. But in the last year it’s become inescapable, in the press, our boardrooms and our homes.

Online dating, movie recommendations, treatment choices, mortgage approvals, map directions, online ads… it may be lurking beneath the interface of our modern lives, but AI’s impact is undeniable. One must ask the question, “Are we prompting the machine or is it prompting us?”

ChatGPT provoked fear and wonder when the Large Language Model (LLM) crashed onto the scene. Some jumped at the chance to produce easy content, while others lamented the possibility that machines could strip the humanity from art. It’s still wonky, but significantly less so in its 4th iteration. The pace of development alone is cause for us to sit up and take note of its disruptive potential.

We’ve seen LLM use span industries and sectors in use cases from chatbots, content generation and even medical research. And as Stijn Christiaens, Co-Founder of Collibra, suggests, “In November of last year, you could actually call it a BC and an AD moment. Before November, AI used to be the data scientists’ purview, and after November of last year, any 12-year-old kid who could speak English could do AI.”

It’s a wake-up call across the board as AI models progress at such speed that common processes that have always been done manually are now being automated using LLMs. But we must remember that though AI has been around since the 1960s, parts of this technology are still new and they’re not infallible. As such, the laws and policies surrounding data governance are more important than ever with a Blueprint for an AI Bill of Rights in the US and the EU’s freshly minted legislation on AI.

Kyle Winterbottom, Founder and CEO of Orbition, quotes Lyndsay Weir, saying “you don’t have to be the first, you just have to be best,” highlighting that, like the fabled tortoise, slow and steady may well win the AI race. 2023 was the starter pistol for AI entering the mainstream.

You don’t have to be the first, you just have to be the best

How Data & Analytics teams are valued is changing

Aside from the obvious use cases of AI across most sectors, its advent has begun to affect the structure of industries, the way businesses are investing, and how data analytics teams are viewed. As long as context is required for analysis and decision making, humans are still in the loop. After all, even if an LLM can turn dumb questions into smart queries, the output quality is relative.

People remain responsible for errors at the hand of AI – think racial prejudice from biased training models impacting job applications, insurance premiums and even self-driving cars. Internal structuring is being shifted to ensure there are trained professionals at hand to challenge the data outputs, tweak the inputs, and mitigate data bias as soon as it becomes apparent. The pendulum is swinging counter to the hype now that Data Governance is the topic at the grown-ups’ table.

Data & Analytics teams have been forced to prove their value, whether that’s financial or otherwise, as the economic downturn forces organisations to re-evaluate spending. Kyle Winterbottom states, “In any uncertain economic times, anything that is deemed to be a cost centre that cannot articulate the value that it’s adding is going to be at risk of being cut.” Whilst inflation is coming down, the macroeconomic outlook affects decision-makers moods. Don’t end up in the red column if you can help it – learn the language of pounds and pennies and you’ll be just fine.

However, while some may want to reduce staff as much as possible in favour of automated AI, it’s clear in some cases that they’re just not ready to implement it. Data management comes first unless you want wonky outputs eroding the faith in shiny AI.

Data management comes first unless you want wonky outputs eroding the faith in shiny AI

Data Products and monetisation

So, Data & Analytics teams need to look forward and prove their value, lean into the new paradigm, and show that they’re the right people for the job even as it’s evolving. Even CDOs face steeper learning curves and the challenge of showing the value of the process behind the function. The differentiation between ‘output’ and ‘outcome’ has never been more relevant.

Data itself has been under the microscope along with the teams that handle it – it’s at the heart of the year’s evolution. But the way we articulate and place value on data also seems to have shifted from cold, hard cash to the more nuanced outputs. “That’s where a lot of the hidden value is. It’s in time saved or errors prevented,” says Susan Walsh, The Classification Guru, explaining that it’s not just financial value that should be ascribed to data products, and the importance of these lesser appreciated benefits.

“Data product” as a phrase has seen a surge in use this year but could mean a myriad of things. As Stijn Christiaens says, “It could be a dataset itself, it could be a dashboard, an AI model.” And the trend that took hold in 2023 also seems to be here to stay, with data products mentioned in Gartner’s CDO framework, and in terms of data monetisation. We take the view that Data Products (at least a product management approach to data) are a great way to surface the right data to the right people, but they must not distract from data quality, governance, or the underlying capabilities necessary to run them at scale.

When we talk about monetisation of data, does your mind go immediately to selling the data? Or do you think of repurposing the data within your organisation to maximise use and profitability? Your data could be put to better use internally within a business, generating several tiers of income by leveraging data to improve customer service or analyse buyer trends.

Semantics may seem trivial, but they matter here. Packaging these up and calling them data products pushes us to view, use and market them as actual products, joining two sometimes opposing facets, data science and business, in harmony. You identify business needs, accessibility and how best this product can be used en masse.

Of course, this is all underpinned by the absolute necessity of clean data. Clean data is the difference between your data product becoming gold dust or garbage.

Clean data is the difference between your data product becoming gold dust or garbage

Lay an expert Data Foundation

As we reflect on the past year, it’s clear to see that it’s been a monumental paradigm shift from those gone by. We now have the technology readily available to implement the algorithms dreamt up decades ago, and AI especially is not just a mystifying enigma relegated to the sci-fi writers and data scientists of the world.

As we enter into 2024, enlightened and cautious, data literacy is of paramount importance. Skimp on your data management and you’ll face fines, distrust, and failure to thrive in the fast-paced landscape we’ve been flung into.

Tread carefully and do the prep work. Talk to Ortecha today to find out how to harness the teachings of 2023 and look forward to the year ahead.



We discussed this topic – and much more! – with our fabulous guest speakers in the 6th online event of our Data Leadership Series. Catch up now!