A data-driven business describes a company that uses data to:
- Make better decisions.
- Improve business processes.
- Gain more customers.
- Keep customers for longer, and, as a result;
- Increase revenues and profits.
Unless you’re running a data-driven business, your staff probably use gut-feel, intuition, emotions, or external pressures to drive decisions. This leads to an over-reliance on the skills and experience of your team. Processes and outcomes will not be easily repeatable, because your staff change over time. In addition, by using data to drive decisions, politics and personality can be swapped for rational certainty.
56% of companies say “data-driven” is just a slogan in their company. This leads to frustration when trying to make use of data. If this sounds like your company, read on.
What is a data-driven business?
A recent survey by EY highlights that 81% of businesses agree that data should be at the heart of all decision making. The same report shows that only 31% of companies have significantly restructured their operations to do this.
So what does it mean to be a data-driven business? And how can you transform your own business to make data-driven decisions and dominate your competitors?
Data-driven businesses are those that make their decisions using data and analytics. If you use data to drive outcomes your business wants, then decisions leading to these outcomes will be made using facts rather than gut feel. So far, so simple.
Key traits of a data-driven business
Some key behavioural traits of data-driven businesses include:
- Executive strategy is based on observable facts and trends, rather than gut feel.
- Decision-makers act on facts first and opinion second, meaning;
- Staff show a willingness to learn from experience and admit when their first judgment was wrong.
- The organisation adopts a “test-first” mindset in the release of new products. You document the assumptions you’re making as hypotheses and then seek data to prove or refute your assumptions.
- You can create an enterprise-wide view of the facts that are not siloed or blinkered.
- Staff are thirsty for information, constantly seeking new knowledge and aiming to put it to use.
- The business prioritises continuous improvement. This is necessary because the road to being data-driven is long and bumpy.
You’ll notice that we have left out any mention of skills in analytics. This is a deliberate omission. The level of analytical skill your organisation needs depends on the decisions that you must make. The more sophisticated your requirements, the deeper the skills in analytics you require. The key point is that you should not start with analytics and work backwards. Start with the decisions and work from there instead.
Too many articles focus on the latest methods to combine, display and represent data rather than on the mindset you need to be able to use data effectively. Data Literacy is a critical capability to build a data-driven business. Unless your staff are willing and able to use data to improve their work any investment in analytics will be wasted.
Why is it hard to be data-driven?
The complexity and scale of most enterprises today creates a major challenge. Driving a small business forward with data is far simpler than embedding a data-first culture across a multinational organisation. For executives in your business to make decisions using data, they first need to know it’s there. Once they know the data is available, the next step is to make sense of it all. The volume of data today makes this task overwhelming for many.
Many organisations face a cultural challenge when changing data behaviours. If your executive team has a history of ignoring facts or distrusts the information provided in reports, then you’re in for an uphill battle. To successfully leverage the full potential of data, your executive team must buy in to the possibilities that success will bring. This demands that we first understand the strategic priorities of the executive team before we can paint a picture of what life will be like in our data-driven business in future.
Challenges for data professionals
For the data professional, the challenge is to work out what decisions your Executives need to make. What outcome are they seeking, and which business questions do they need to answer to reach these outcomes? To build a data-driven business you need:
- To provide the right data, and;
- Make it available in the right place,
- At the right time,
- With trust in the quality, and;
- Consistency in meaning across enterprise silos.
This is a major challenge. To deliver on the promise of the bullets listed above, you first need to understand the overall strategic direction of the company. Secondly, you can unpick the major decisions that must be made. Lastly, you can start to identify the data needed to support these decisions. Depending on the organisation you work in, you might need to build a data governance function to then ensure that data is fit-for-purpose through time.
What are the benefits of being data-driven?
Companies in the top third of their industry in the use of data-driven decision making were, on average, 5% more productive and 6% more profitable than their competitors.
“Insights driven” public companies grow at 27% year on year.
Using customer analytics heavily makes you more likely to outperform your competitors, according to McKinsey. Being data-driven makes you 23x more likely to acquire customers, 6x more likely to retain those customers and 19x more likely to be profitable as a result.
The gap between the leaders and the laggards is wide and continues to grow. High performing organisations say their data and analytics investments have contributed 20% to earnings over the past three years.
All this points to a high return on investment from data activities. If you can get this right, your organisation stands to win big.
Are there any drawbacks to being data-driven?
You’d think given the benefits of being a data-driven business that it’s all plain sailing. Unfortunately, this is not the case. Companies get drawn into a bear pit of wasted investment chasing the “blue-sky” dreams from major consultancies. As a result, companies are throwing money at data and analytics without a clear line of sight into their expected Return on Investment. Just because business leaders like Amazon and Facebook are winning with data, does not immediately mean you’ll follow suit.
Gartner Analyst Nick Heudecker estimated in 2017 that 85% of big data initiatives had failed. Less than half of the firms that responded to a Harvard Business Review survey thought their analytics projects had delivered measurable results. Another Gartner analyst has a dim view of the future outlook too:
Through 2022, only 20% of analytic insights will deliver business outcomesAndrew White, Gartner
The bulk of the failed projects are driven by hype. Vendors sell the benefits of being data-driven without being honest about the challenges too. This leads to a clamour to “keep up with the Joneses” and the firms that are technology laggards end up paying the price. To avoid these traps, you need to document great requirements and set yourself up to succeed. Do not start with Technology.
How data-driven are businesses today?
NewVantage Partners have been surveying executives across businesses for the past decade. The most recent survey results have a rather disappointing theme. Despite heavy investments, companies are still struggling to make the progress they seek:
- Less than half (48.5%) are driving innovation with data.
- Only 41.2% are competing on analytics.
- Only 39.3% are managing data as a business asset.
- Less than a third (30.0%) have a well-articulated data strategy for their company.
- Based on Cognopia’s in-house research, 32% of companies state that a data strategy is in place. Many of these documents are not well articulated and are poorly received by the business.
- Only 24.4% have forged a data culture.
- And only 24.0% say they have created a data-driven organisation.
If your organisation lacks a data governance function today, you will struggle to unlock the value in your organisation’s data. Your data is likely to be disorganised and not fit for more advanced purposes unless you’ve actively managed that data for those purposes.
How to build a data-driven business
Now we have covered what it means, let’s take a look at how you can become data-driven yourself. The obvious goal is to improve business performance. Unless we can deliver a performance boost, we should not invest in data activities. Let’s take a look at how to turn this into reality:
How can you use data to improve business performance?
Businesses that are data leaders are clearly delivering impressive results. We have identified 9 ways these firms are pushing the envelope and using data to improve their businesses:
- Using data to improve decision-making accuracy and speed.
- Gaining control over cost and spending metrics using data.
- Understanding their customers better, and drive increased loyalty and spending.
- Identifying new markets or growth opportunities.
- Reducing fraud and eliminating other business risks.
- Creating better products and services.
- Selling or trading their data with partners and customers to generate new revenue streams.
- Increasing employee engagement and satisfaction.
- Improve efficiency across critical business processes.
What makes this challenging is that you need to translate these high-level data ambitions into a practical, pragmatic data strategy that fits your business today.
Create a data strategy aligned to your organisational strategy
To do this, you must identify the bullet point above that will benefit your business the most. Your Annual Report will show where your business is heading. If you’re in a private company that does not publish Annual Reports, seek out strategy documents that indicate what your key objectives will be.
Once you have these documents, you can start the work to align data objectives with business objectives.
The alignment process, in steps:
- Firstly, find a strategic objective that aligns with one of the bullets listed above.
- Perhaps your firm is trying to expand into a new geography – in which case, you’ll pick “Identifying new markets or growth opportunities”.
- Second, brainstorm ways to use data that will help meet this strategic objective.
- Continuing the example, you may have data on hand showing the investment needed to open a similar geography previously.
- You might also consider data you do not currently have, such as GDP or population demographic data. This data could help your executives make the choice to launch.
- Lastly, you could look for partner data. Do any of your partners already operate in the geography you’re targeting? Perhaps they would share that data with your firm, or exchange it for data you have that is valuable.
- Finally, engage with the executive responsible for making the decision.
- Tell them about your plan to help them.
- Explain how you want to provide more data to help them make a better decision and eliminate some of the doubt and uncertainty they are facing.
- Ask them what data they would like if they lived in a perfect world.
By aligning your data activities against the strategic goals of your business, you’ll get more enthusiasm from the executive team. Instead of asking the Execs to help you put data first, show them how you’re putting their decision first.
Data-driven decision making made easy
The first of our 9 ways to use data to improve business performance focuses on decision making. In reality, improving decision making will lead to improvements in each of the subsequent 8 bullets. As a result, we want to take time to focus on this step and walk through how to make this a reality in your Business.
Data-Driven Decision Making (DDDM) is straightforward. The aim is to make better decisions, more quickly using the data we have on hand. Ask any executive how confident they are in major decisions and you will find (privately) that many are forced to decide without the full picture available. This is because it is hard to provide perfect information at precisely the right time. In addition, many executives will also tell you that they lack trust in the data they’re presented.
Less than half of companies say that data is highly valued for decision making. We have to fix this.
Do not just rush off to grab data around the business. Firstly we have to focus on the data that matters. In our opinion, the best approach to identify this data is to use Decision Intelligence.
Using Decision Intelligence to identify Critical Data
Decision Intelligence is a new technology that helps us model complex decisions. As a result, it’s perfect for Data-Driven Decision Making. Proponents such as Cassie Kozyrkov (of Google) insist that it’s a vital science for the AI era, and we agree. In addition, we think the core principles can help any organisation make better decisions.
Lorien Pratt’s book “Link” is a fantastic primer on the topic. The second chapter introduces “Causal Decision Diagrams” (CDD). The fourth chapter explains how to run a workshop to create one of your own. The main reason to create a CDD is to model a complex decision that is critical to your organisation’s growth or survival.
How to create a Causal Decision Diagram
The first step to creating a CDD involves identifying the decision you need to make. If you need help to pick an appropriate business decision, click here. Next, you need to get the support of an executive sponsor. Even if they don’t participate in the full CDD process, getting their guidance will make a huge difference. You will need to define a problem statement that the CDD will overcome. Get the executive sponsor to sign this off.
An example decision could be “We are looking to grow the business by 10% in 2021. Should we enter the Chinese market with our technology?”.
Such a decision is inherently complex. As a result, you’ll want to bring together a diverse team with experience that can help make sense of it. The more diverse the team, the better.
Your executive sponsor may need to lay down some ground rules. For example, there might be a maximum budget of $1m available to support the entry into a new market. Unless you knew this before you began modelling, you might create ideas that will never be used. Therefore you should jot all these down and use them to frame the modelling session.
Start building the CDD
The first step in building out the CDD is to identify the outcomes your team wants to achieve. In our case, one of them is stated in the problem – the desire to grow the business by 10% in 2021. Awesome! Next, we need to flesh this out and document other outcomes. For example, we might have an obligation to hit specific environmental targets that are laid down by law. This needs to be documented, as the decision to enter a new market might have an impact on such targets. Finally, there may be other outcomes of importance too, such as maintaining profit margins or protecting Intellectual Property. List them all.
Outcomes alone are not good enough. Once you’ve brainstormed as many as possible, it’s time to get specific. Start by defining measurable goals that you can achieve. Our first example “grow the business by 10% in 2021” is an example of a measurable goal. We’ll know if we achieve it. Define the metrics by which success will be measured.
Flesh out the details
Now you know where you’re going, you can start to document how you’re going to get there. Time for another brainstorming session. In this session your team should list as many levers as possible – options you are in control of and can select that might get you there. For our example, this might include a choice of pricing for the new market. Maybe we need to consider shipping methods or whether to open a factory in that market. List all the potential options that could be used to help.
Lastly, you’ll need to identify “Externals”. These are things beyond your control that may impact your decision. Covid-19 is an example of an External change that could dramatically impact your decision. Whilst it’s not possible to predict them all, your team can list as many as possible in this session.
Stitching the CDD together
Once you’ve got a great list of Levers, Externals, Outcomes and Measurable Goals you can begin to stitch them together. These are the links mentioned in Lorien’s book. This is when your team must use their experience and knowledge to bring the Causal Decision Diagram together. For example, one of your team members might know the competitor’s price point in China is $50. Your own price point to manufacture the goods might be $30. You then have a lever to choose whether to ship products from your existing plants or a lever to set up a new plant in China to supply the goods locally.
As you consider these two levers, you need to link them to the outcomes and goals your team defined. Shipping goods in from an existing plant may cost $10 per unit, meaning you can make a 20% profit margin (above your goal). At the same time, the carbon footprint of shipping may negatively impact your environmental targets. This might take this option off the table.
Carefully link each lever with the goals, and pay attention to any feedback loops or unexpected interactions.
Identifying the Critical Data
In our experience, creating the CDD alone has huge value for your business. Most enterprises lack a coherent map of the major decisions they make. Silos exist between teams that may be competing with one another for the same resources, or worse, pulling the company in 2 different directions.
As you go through the links in the chain there will be areas where you have uncertainty. There will be other areas where you have solid data that points to a specific choice for that specific lever. In both cases, this describes the data that will be useful to your executives when making this important decision.
Take the shipping lever as an example. As you analyse this, you may find that you need more data on the carbon footprint of different methods of transportation. However, once you’ve investigated the lever around setting up a plant in China, it might be obvious that there’s no way you can produce, ship and sell products from your overseas plants more cheaply than if you made them locally. As a result, there’s no point in looking for data on the carbon footprint. It is irrelevant to your decision because you would never choose to use that lever in the first place.
By using tools like Decision Intelligence you can identify the areas where Critical Data resides. This will help you avoid the trap of “over-improving” data that is not used by the business. This will save you time, money and resources so you can focus your efforts on data that delivers meaningful value to your executive stakeholders
Pulling it all together
Making better decisions more quickly is the main reason to become data-driven in the first place. What’s crucial is identifying the important data to help guide decisions. Decision Intelligence is one technology that you can use to map this out. Even if you choose not to leverage DI, the key takeaways from this section should be:
- Focus on important decisions that add real value to your business bottom line. By doing this, you’ll increase your senior executives’ enthusiasm to support the project.
- Identify the Critical Data that supports these decisions. Once you’ve identified this data you can then ensure it’s fit for the intended purpose.
- Do not aim to make all the data available to everyone. This is “boiling the ocean” and will waste your project time and resources.
Start slowly and build on successes. You should take every opportunity to document and demonstrate the value your data-driven approach brings. Doing your homework upfront allows you to clearly demonstrate a strong Return on Investment. Subsequently, you’ll find it much easier to persuade other executives to follow your approach.
Challenges to becoming a data-driven business
As you can see from the sections above, there are many reasons why businesses struggle to become data-driven. Let’s take a look at some of the major challenges and explore how to overcome them.
Data is poorly governed
A data governance function exists to ensure data is fit for purpose. This function acts to exert executive authority over business data. For example, the term “Customer” must be defined clearly:
- Does it mean anyone that has ever purchased something from your business?
- Could it mean anyone that bought something in the last 12 months?
- Or is it someone with an active subscription to an ongoing service?
As you can see a simple term like “Customer” can have many meanings to different departments. To cut through this noise and make sense of our data, we first need to document what it means. After that we can start to dictate data quality rules – do we need email addresses in our customer records? How often should we validate that these email addresses are active? Who should make these decisions in our business?
When data governance is absent, data falls into disarray. Making sense out of badly managed data is an uphill battle. Reports based on badly managed data erode trust from your executives and create disagreements. You have to fix this before you can become data-driven.
The state of data maturity in 2021
Cognopia has been delivering data maturity assessments for free since early 2020. The results of these assessments have given us an unparalleled insight into the state of data maturity across Industries and Geography. This, in turn, has led to the comprehensive report on data maturity we link to below.
Your data maturity matters because it tells you what skills and capabilities your business has today in managing and using data. Unless you know this you cannot plan a new data initiative. We’ve seen countless projects start and fail because they were too ambitious and bit off more than they could chew. Do not do that.
Our data maturity assessment grades the company on 9 key characteristics. At the same time, it rates the company across 5 maturity levels. This allows us to provide very specific guidance on how to improve.
One major trend we have seen is the aspirations of companies taking our quiz is high. In many ways, this is a good thing. As Splunk found out, 90% of firms believe that every organization must extract value from data to be successful in the future. Consequently, we see a lot of ambition amongst our respondents.
On average, the scores to “aspirational” questions are 3.07 (out of 5) whereas the “operational” question scores are 2.27. That’s a 26% difference! This shows that companies say the right things, but they are yet to start doing them.
Key findings of Cognopia’s 2021 State of Data Maturity Report:
We discovered a range of challenges in our 2021 survey. A summary of these finding is below:
- 56% of our survey respondents say data is a strategic enabler or a competitive differentiator for their business
- However, 35% of firms are only looking at data value in terms of cost, unable to see the bigger picture
- Just 32% of firms have a data strategy today, the leaders have data maturity scores substantially greater than those without any strategic direction
- Senior stakeholders are yet to embrace data governance – only 16% of firms had active participation from their most senior resources
- A third of businesses (33%) had no dedicated budget to improve their data
- Data Ownership remains a murky and confusing topic – 42% of firms had no data owners or relied on I.T. to play this pivotal role
- 37% of firms knew their data was poor quality, yet still had no process in place to improve it
- Metadata management tools are yet to hit prime-time – just 2% of participants report wide usage and adoption of these technologies
- Data teams fail to set and measure KPIs for their own work – 77% of our respondents are unable to tell us whether things are improving as a result of their work
Common data maturity problems
The first major data maturity problem we see is a lack of a Business Case. Respondents tell us that they have no business case, or that their business case is poorly formed. In turn, this leads to a lack of executive engagement in projects and under-funding for data initiatives. You have to build a good Business Case if you want your data-driven initiative to last for the long term.
Secondly, we often see a challenge in Data Ownership. Put simply, a Data Owner must be accountable for the data they “own”. This role is critical. Being accountable for data means you need to have the power, budget, and influence to improve that data. Specifically, this role has to have the authority to “do something” about bad data. Usually, this means funding an improvement project, appointing “Data Stewards” that do the improvement work, and clearing roadblocks internally that cause the bad data in the first place.
Common Data Ownership problems
- No one owns the data. This means there’s no accountability for data in the organisation. Without accountability, there’s no one to report data issues and errors to.
- 16.3% of firms we surveyed had no one accountable for their data.
- IT owns the data. This is not much better than the bullet above. IT might own the system the data resides on, but they do not have intimate knowledge of the data within that system. As such they cannot be accountable for data issues.
- 25.6% of firms we spoke to were making this mistake
- The wrong person owns the data. Data Ownership can be a tricky subject. Just because a business sponsor has been appointed does not automatically make that the right person for the role.
- 18.6% of firms we spoke with were in the process of appointing Data Owners. This meant we were able to help guide their choice and support this key role.
Other problems that exist include people “hoarding data”. Unfortunately one side effect of calling this role a “Data Owner” can mean some people then believe they get to do whatever they like with it. This is not the case and needs to be nipped in the bud quickly.
Data quality problems
Data quality is a big deal. If your data is poor quality, your decisions made from that data will be poor quality too. As such, achieving high data quality scores is an essential step in building a data-driven business.
Understanding data quality requires a post of its own. In short, the data must be fit for its intended purpose. This means talking to the business, understanding the decision, and agreeing on the expected quality metrics. Get this right and you’ll delight your Executives and make everything else easier.
Our research discovered this was a hard challenge for our respondents. 37.2% of the firms we surveyed stated that “Data is viewed as poor quality, but no formal processes are in place to fix it”. Another 27.9% of respondents said that IT was responsible for fixing bad data.
As with Data Ownership, IT cannot be responsible for fixing poor data unless they know what “fit-for-purpose” means. In addition, fixing bad data gets expensive. Eliminate the root cause of bad data. To do this you must change the way your data is captured. This will lead to long term success and reduce the cost of managing data.
In our 2021 poll, 53.5% of respondents said their executive leaders were either not interested in data governance, or that they had limited interest in the subject. As a result, there were no budgets for improving data and the projects that did exist were poorly sponsored.
We are not the first organisation to report this. Max Henrion‘s article “Why do so many Big Data Projects Fail?” has this to say:
- Management resistance and internal politics (reported by Gartner).
- Insufficient organizational alignment, lack of middle management adoption and understanding and business resistance (HBR)
Changing the culture of an organisation is a challenge. The only way you will make progress is by demonstrating tangible business value. Unless you can demonstrate a return on investment you will never get your management to care. Consequently, data practitioners need to stop blaming the management for these failures and start to learn how to communicate their value more effectively.
Data literacy is lacking
According to Qlik: “Data literacy is the ability to read, work with, analyze and communicate with data”. This means that the decision-makers need these skills before they can use the data you’re providing them. Indeed, you can provide the best analysis in the world, but unless your executives can understand it you have wasted your time.
Similarly to our comment on data analysis, start by understanding the level of data literacy you have today. Next, consider what level of data literacy you’ll need tomorrow. Train your audience according to their needs, otherwise, you’ll bore people and put them off data for life.
To do this effectively, you must tailor your training to your audience. Some people want to “geek out” on the details, others are happy to accept the results from your “black box”. As such, train those that need to be trained. As long as there is trust in the results you provide, and the data is acted upon, you have succeeded.
Hopefully, this guide has inspired you to use data more productively in your business. Start small, pick achievable outcomes, and you’ll be well on your way to join the trendsetters.
If you have any further questions about how to implement any recommendation we have made, please get in touch. We love helping people get more value from their data.
No question is too much of a problem. Feel free to drop our team a line below and we will help you embed data-driven behaviours in your business.