Data-driven decision making with AI: everything you need to knowBy Jon Taylor on April 22, 2021 - 10 Minute Read
Data is now playing an increasingly-prevalent role in our lives, impacting everything from everyday decisions to how we run our businesses.
With companies now relying on data driven decision making to reach more customers and with algorithms funnelling people to the content they want to see, data is now playing a key role in determining a lot of the information we see online. This data gold rush is an exciting time for businesses. If you manage your data sets correctly, and leverage your data to its full potential, it can be used to make great business decisions in a number of key areas.
In this article, we’re going to take a deeper dive into:
- What is data-driven decision making?
- Why data-driven decision making is a game-changer for business
- Data-driven vs. Decision Intelligence
- How do you make a data-driven decision?
- Potential problems with data-driven decision making
- How can AI improve data decision making?
Let’s get started.
What is data-driven decision making?
Data-driven decision making is when businesses turn to data sets, insights and patterns to help them make better decisions based on facts and figures instead of gut instinct.
At its core, data-driven decision making allows brands to figure out what’s working well by measuring Key Performance Indicators (KPIs) and metrics while tracking if their goals are being hit based on verified data points.
With the sheer amount of data businesses now collect (think about every time you’ve been asked for your email address or job title), analyzing it all and using it to make smarter decisions can be complex and incredibly time-consuming.
Brands are now spending trillions to digitally transform their businesses and take advantage of datasets, yet research from Gartner suggests that 85% of data projects fail. Even big brands like Adidas have struggled to switch from traditional decision making to a process powered by data – admitting their decisions around digital advertising were a failure.
We had an understanding that it was digital advertising – desktop and mobile – that was driving those sales and as a consequence we were over-investing in that area. We had a problem that we were focusing on the wrong metrics, the short-term, because we have fiduciary responsibility to shareholders.
Global Media Director, Adidas
The good news is that with the right tools and planning, you can avoid the mistakes made by Adidas and use data-driven decision making to supercharge your processes and power the right commercial decision, all the time.
Why data-driven decision making is a game-changer for business
It’s not just that data can give you a more accurate view of how your company is performing than manual analysis. Using data also helps eliminate one of the most problematic areas of decision making: cognitive bias.
There’s no escaping the cognitive biases we humans have – after all, there are over 180 of them. These biases cloud our judgment and make it harder to make decisions based on facts, which is toxic when trying to make good decisions for business.
That’s where data-driven decision making can be a game-changer for your business.
It acts as a mirror of truth: it doesn’t tell you what you want to hear. Instead, it eliminates cognitive biases and highlights the best path forward for your company based on metrics and KPIs – not emotion. And although this may sound like another form of decision making – Decision Intelligence – there are some key differences you should be aware of.
Data-driven decision making vs Decision Intelligence
As we’ve explained, data-driven decision making is rooted in data – so how is Decision Intelligence any different?
The sheer amount of data Decision Intelligence can analyze and its ability to make accurate predictions sets it apart. While data-driven decision making is rooted in metrics and KPIs, Decision Intelligence is powered by artificial intelligence (AI).
It’s only recently that this technology has become available to businesses instead of being reserved for mega-corporations with huge data science teams. With AI now more accessible, brands have a chance to leverage data from anywhere (no matter what state it’s in) and eliminate bias and silos from their workflows to predict customer behavior and build better supply chains.
Enough about how exciting all of this is – let’s look at how you can implement it into your own workflows 👇
How do you make a data-driven decision?
There are lots of different ways you can make data-driven decisions – from changing the way you sell products to shaking up your management style.
A great example of a company making better decisions using big data sets is – you guessed it! – Google.
The sheer amount of data Google collects is mind-boggling: it processes 3.5 billion searches a day, with each one triggering request queries across 20 billion web pages. But the company has also turned data into an internal superpower to study how well its team is performing.
- Google collects qualitative data from employee surveys and performance reviews to see how well their managers were performing. By gathering data sets across quartiles, the company’s data scientists were able to examine which manager had the best team productivity and morale as well as turnover and revenue
- This data was then used to determine which of Google’s managers were generating the positive results expected of them and who needed further training
- Google then introduced its “Great Managers Award” program, which continuously collects data from employees and metrics to uncover the behaviors of their best-performing managers. The findings are then used to revise and improve the company’s management training
Google is a perfect example of how data-driven decision making can overcome cognitive biases. While managers may think they know how to get great results from their team and what behaviors work, collecting data actually helps them find ways to get better results.
Here’s how to make data-driven decisions like Google 👇
Define your decision making goals
Before getting into how to make a data-driven decision, let’s start with why.
Are you making decisions for a marketing campaign, or are you trying to increase sales or stop churn? The type of decision you are making will determine what data sets you’ll need to look at, but the research shows no matter what decision you’re making, the more data, the better.
Think about how tying data-driven decision making to your business goals will help achieve them. You may want to follow in Google’s footsteps and use data sets to improve your management processes. Or maybe tracking how effective a brand awareness campaign is performing will help you decide whether or not you need to look for new channels to distribute your content.
No matter what goal you want data to help you achieve, think about how effective data will actually be in the process.
No matter what goal you want data to help you achieve, think about how effective data will actually be in the process.
Analyze the results and make some (data-driven) decisions 🤔
With big data sets, it’s easy to get lost in all of the numbers, and it can be very difficult – if not impossible – for humans alone to number-crunch to the level required in order to make a data-driven decision.
As we’ll cover in more detail shortly, technology like AI can help you break down complex data sets so you can analyze your data at pace and scale.
When Marshalls, a leading hard landscaping manufacturer in the UK, decided it wanted to take an (unbiased) look at its processes, the company turned to data-driven decision making to help. Because its business model relies on multiple contractors bidding on projects, the company’s reputation relies on keeping the process quick and efficient so contractors know if they’ve won a project or not.
The company’s large data sets meant analyzing them manually was out of the question. So, Marshalls leveraged AI and machine learning to unify its data. Marshalls turned to Peak and used its Decision Intelligence platform to gather data from multiple datasets and systems to get a better idea of how its processes were performing. The business then had access to:
- Real-time insights that were integrated into existing ERP systems
- Better visibility on the open projects and quotes as well as average number of sales per customer/per month
- Securely stored data in Amazon Redshift along with datasets in Amazon S3 and Amazon SageMaker to deploy the endpoints
Using data has not only given Marshalls’ customers a better experience – it’s been good for the company’s bottom line, too.
Thanks to the power of AI and data, customers are able to get the answers they need in terms of bid decisions and quotes quickly and efficiently, significantly speeding up our sales cycle.
AI Solutions Architect, Marshalls
Potential problems with data-driven decision making
Data is changing the way businesses make decisions – but that doesn’t mean outsourcing every decision is the way to go. Businesses can run into problems when they rely entirely on data dashboards and analytics to call the shots – or when they ignore them altogether. Watch out for these potential problems associated with data-driven decision making.
Letting your cognitive biases take over
We’ve already touched on the danger of cognitive biases in the decision making process, but to avoid it happening, you need to know what to look out for.
When assumptions are given the same weight as data, cognitive biases can dominate the decision making process. In business, we see this happening in a couple of ways:
- Embracing group think: Cognitive biases squash dissenting opinions. This creates an environment where people aren’t comfortable accepting new data leads to “group think,” where the majority opinion rules
- The “yes” man: Instead of listening to data and considering changes, those in charge only accept results that conform to their own beliefs – even when they’re wrong
- Sticking to old habits: When new technology is introduced, some business leaders escape to their comfort zones, believing traditional processes are still more beneficial. If this happens, it’s harder for a business to embrace new data and metrics to make better decisions
The first step to overcoming biased behavior is recognizing it. If you’ve found yourself acting like any of the scenarios we’ve just talked about, there’s a chance you’re sabotaging your own data.
Thinking you know more than the data
We’ve all been there – no matter what the data is telling us, our gut is telling us something different. Trusting your instincts over cold, hard data is another place where decision makers get into trouble.
Suppose a quarterly sales report has come in, and it shows that some of the team’s sales reps are consistently underperforming. A manager’s instinct may be to explain it away, even if the data is there and the results are clear. This situation also happens when we’re emotionally invested in a marketing campaign or customer retention effort – even if the data tells us it’s failing, we instinctively push back. No matter how hard it is, it’s always better for your business to trust what the data says.
Neglecting the all-important human touch
There are some circumstances where decisions need a human touch. Let’s say your data sets are showing that your team’s calendars are only filled to 80% capacity. But by ramping them up to 90%, you’ll be able to add more projects to your pipeline and earn more revenue.
What the data sets don’t tell you is that this would be catastrophic for your team. Although your team may work at 80% capacity on paper, the data fails to show that the remaining 20% of their time is spent answering emails, communicating with customers and collaborating with their co-workers.
Imagine what adding another 10% onto their workload would do? Total team burnout. That’s why human judgment mixed with data is the ideal recipe for making decisions – and that’s where AI comes into play.
How can AI improve data decision making and avoid these problems?
Combining AI and humans to make better decisions works so well because both are very smart processors.
Humans bring a rational and pragmatic approach to the decision making process, whereas AI analyzes data and makes it harder for cognitive biases to dominate. And to leverage the full benefit of AI, businesses need to move from making data-driven decisions in their traditional sense and instead embrace workflows powered by AI.
On the surface, “data-driven” and “AI-driven” may not seem all that different, but AI decision making goes one step further than data-driven decision making. Instead of just pulling data and combining it into dashboards, AI processes it, extracts insights, runs multiple scenarios and makes predictions and categorizations around outcomes.
Integrating AI workflows is no longer just for mega-corporations and data scientists—businesses big and small are now discovering just how valuable their databases are. From training better managers to streamlining processes and building better systems for your customers, the benefits AI can add to your workflows are endless.
Here at Peak, we’ve helped companies like PrettyLittleThing, Speedy, Aludium, PepsiCo, AO and many more discover the power of using AI to make smarter decisions. We’d love you to read more about their success.