How to Be More Data-Driven Manager in 3 Easy Steps
You only have to be as smart as a 6th-grader
What if you need more data-driven decisions, but neither the boss nor the employee knows what that means? I know a lot of people are told by their managers to be more data-driven. Sometimes, it’s because the boss was told by their boss. Or worse, they read on social media that all the other companies were becoming more data-driven and decided it sounded like a good idea.
Since most of them have no idea what that means, why they should be more data-driven, or how to do it, they can’t help you. You decided to figure it out on your own, and that landed you here.
No worries. I will break down the three keys you need to be more data-driven. Only one of them includes math, and I promise to keep it simple. I will show how it can be much simpler with AI than complex spreadsheets, which can be daunting for many.
Step 1: Change your mindset
You need to think about when it’s required. This is the hardest step of all. It is also the most impactful because it defines the “why” you want to become more data-driven.
How many conversations have you had with your team that sounds like this:
Boss: We aren’t selling enough.
Team: We ask for the sale on every call. We don’t know what else to do.
Boss: You need more training. I’ll talk to Human Resources.
Team: Whatever. Can we go to lunch now?
Training alone is never the right answer.
When you use or hear qualitative words, it’s a trigger to go into data mode. For example, words like “enough” (ex: we aren’t selling enough), “better” (ex: we need to provide better service), and “decrease” (ex: decrease our expenses, decrease time on the phone, decrease complaints).
These amorphous terms create unnecessary friction and don’t help you drive improvements. They make everyone feel bad that they aren’t “better” and turn any real problem-solving attempt into a pity party.
Let’s stick with the selling example. If you want to sell “more” (trigger word), how much is more? How much are you selling now? Regardless of your company or product, there are specific steps to making a sale. What are the current numbers at each step (ex: incoming or outgoing calls, first or second conversation, present the solution and close the sale).
By looking at the numbers on each step, now you can have a better conversation about which step needs to be “better”. By knowing where you are starting and where you want to be, you generate more creative ideas from everyone on how to close the gap.
When you first start thinking like this, you will often find that you don’t have the data you need. But if the number you want is the result of an action, the data exists somewhere. If a computer or human answered the phone, there’s a data point for it. If someone or something typed the answer into the system somewhere, that data exists somewhere.
If you still can’t find the data? Now you have something actionable to take to your manager, who should be able to help you find the data or a similar piece of data. And if they can’t, well, send them to me. I’ll teach them.
Step 1 Takeaway: Any time you say “I wish we were better” or “this needs to be different”, try to attach a number to it.
Step 2: Do the math
Wait! Don’t run off yet.
You only need to understand three basic math concepts to be data-driven. Tools like spreadsheets and artificial intelligence can help.
First, we will look at the tools and then go through each of the three main math concepts you’ll need:
- Counting
- Percentages
- Comparisons (less than, greater than, equal to)
Tool options: Get a little assist from AI
For those of you who are comfortable with spreadsheets, they are still an excellent tool for data analysis. They allow you to quickly collect, review, analyze, and summarize large amounts of information.
Learn how to use filters, a few formulas, and pivot tables, and you’re all set for most analyses. With Microsoft adding AI functionality, including Copilot, you have some AI help available, too. Especially since spreadsheets are free to most corporate employees, you shouldn’t incur any additional costs.
Many managers, though, aren’t as comfortable, and some even use spreadsheets as a reporting tool instead of an analysis tool. For you, artificial intelligence tools will be a tremendous way for you to get into the data game.
Tools like ChatGPT or Google’s Gemini are simple to use and able to handle most basic analytic functions. I’ll use ChatGPT in later examples, simply because that’s the one tool that most people seem to have tried.
Even if you are comfortable with spreadsheets, they can unnecessarily complicate the analysis. Adding an AI tool on the side will greatly improve the time and quality of your analysis.
And for those of you looking for more than a spreadsheet, tools like Tableau, PowerBI, PolymerSearch, and Alteryx have AI helpers either available or coming soon.
Counting
Data is just a collection of things, and you probably learned early on how to count things and distinguish when one thing wasn’t like another. That really is the basis of all data analysis. Once you find your data, count it. Then, group it in different ways or find averages of the things to set standards.
How many times did something happen? Tracked over time, this helps you see if you are improving or not, especially if you look right before and right after you make a change.
How many times should it have happened but didn’t? This one helps you decide whether it’s worth making a process change. If something happens 99 times out of the 100 it should happen, don’t upend your entire process to try to capture that last one. But if it only happens 20 times out of the 100 it should, and it's an important thing to happen, it’s worth making a big change to close the gap.
How many time did it start happening and stop before it finished? This is really important for satisfaction. If a process is stopped before it finishes, customers or employees are probably unhappy about something in the process. Or, you lost information that you will need later.
How many times was it done so amazingly that something else happened? (ex: Google reviews, customer referrals, thank you notes) This one is huge for finding moments to praise your employees.
If you want to get fancy, learn how to summarize in a Pivot Table which will count all the things for you very efficiently. You might even throw the whole bunch of data into your favorite AI tool and tell it to count for you. (Warning — do not upload client personal information into AI. Check with your compliance team on your company’s AI policies.)
But if you want to stay simple, just count. The numbers will still be the same. These counts become even more important when they are tracked over time.
Here’s some important things to count:
- Incoming work: The number of things that come into your team (ex, customer calls, month-ends that require reports, help desk requests)
- Production: The number of things that go out of your team (ex, sales from customer calls, projects completed, financial reports)
- Throughput: The [average] time it takes for something to go through your team.
- Aging items: The number of things that have been in your team for longer than they should be (ex, calls that are over 30 minutes,
- Error rates: The number of things that left your team but came back (ex, errors in programming, returns from sales)
Percentages
I know. This one can be intimidating. I failed seventh grade math so I get it. I now have a Bachelor’s in Math because I wouldn’t give up. Let’s make it simple. A percentage is really just a this-compared-to-that number.
This is important because it allows you to compare things that aren’t the same. With AI tools, you really don’t have to know the math details now. But you need to know enough to ask the right questions.
Key formula 1: (How often did it happen) divided by (how often it could have happened). Multiply that by 100 to get the percentage.
- If my team sold 20 products after taking 50 calls, that means they sold 40% of the possible sales. (20/50= 0.4 * 100 = 40%) In sales lingo, that’s called a close rate.
- If I do something to increase my incoming calls, I can use that close rate to predict how much it will increase my sales. Take 40% of whatever you project your incoming calls will be.
- The same formula works for anything. If an employee calls in sick 10 times out of 40 work days, they have a (10/40 = .25 * 100 = 25%)
- Or feed the whole word problem into AI: if we sold 20 items out of 50 calls and I increased the calls to 70, how many should I sell?
Key formula 2: Are we doing better or worse? This is where percentages come in really handy. But rather than do the math, let’s ask ChatGPT to do it for us. I asked:
They only sold 20 today vs 40 last month, so they’re doing worse, right? But the calls are much lower. So is the team doing better or worse? Here’s the answer:
To evaluate if your team is doing better or worse today compared to last month, we will examine the sales conversion rates for each period.
### Step 1: Calculate the Conversion Rate for Each Period
The conversion rate is defined as the number of items sold divided by the number of calls made. It measures the effectiveness of the sales efforts.
**Today’s Conversion Rate:** - Items Sold: 20 - Calls Made: 52 - Conversion Rate \(CR_{Today}\) = \( \frac{20}{52} \)
**Last Month’s Conversion Rate:** - Items Sold: 40 - Calls Made: 114 - Conversion Rate \(CR_{Last Month}\) = \( \frac{40}{114} \)
### Step 2: Perform the Calculations
Let’s calculate the conversion rates for both periods.
### Step 3: Analyze the Results
**Today’s Conversion Rate:** Approximately 38.5%
**Last Month’s Conversion Rate:** Approximately 35.1%
### Conclusion:
Comparing the two periods, today’s conversion rate (38.5%) is higher than last month’s conversion rate (35.1%). This indicates an improvement in the sales efficiency, as a higher proportion of calls are resulting in sales now than in the previous period. Therefore, your team is doing better today compared to last month.
It would have been easy to think the team was slipping. But in reality, the real problem is the calls aren’t coming in. Looking at the data allows you to focus your time on the actual problem.
Step 2 Takeaways: Do the math. Count things, look at the numbers over time, turn numbers into percentages to see if they are improving or deteriorating. Focus on asking the right questions of the data and let AI do the math for you.
Step 3: Evaluate the answers
No one will know that you are now a master of data usage if you don’t communicate it. The steps in your explanation are the same as the steps you took to get here. You just need to make your thinking visible.
- Explain the problem you had to solve. Include the data that shows how big the problem was.
- Explain the end result and what you changed it to. Include the data that shows how big the problem is now.
- Explain the path to get there. What steps did you take, where did you misstep, and how did the numbers change over time from when you first found the problem to now that you’ve fixed it? This is a great time for pictures — a line graph of how the numbers changed over time.
Any time you make a change, you should have the starting data, the data along the way, and the ending data. The actual data, the thing you counted, probably won’t change. It’s the numbers themselves that will change.
How does ChatGPT help here? I asked it to graph the outcome of the earlier conversation so I could show my boss that the sales might be down, but it isn’t because my team is slipping.
I didn’t like the first graph it gave me. It only graphed the conversion rates as bars. That didn’t show the story I wanted to tell, that the problem was the incoming calls, not the team’s performance.
So, I told it to switch the conversion rates to lines and add the raw numbers as bars. It was better, but then I had to switch to having last month on the left, so it was more obvious. There. Now, it’s obvious at a glance that the call volume is down, but the conversion rate and the team’s efforts and abilities are actually up.
Always make sure you include the timeframe and source of the data you used, as well as any assumptions you had to make on the way. You will forget so you are improving your credibility and saving your memory.
Caution
PLEASE don’t distribute spreadsheets with all the data and leave it to your audience to do their own analysis. That just wastes everyone’s time and destroys your credibility. Being data-driven means that you used data to improve something. It does not mean that you have all the data and think everyone else should just see what you do!
Side note
Assumptions are fine as long as they can be reasonably seen to have not significantly impacted the results. For example, you often have other changes happening at the same time you are trying to improve something.
If you are trying to improve the time your team spends actively talking to customers, but the company is also trying to decrease the time employees spend in training, you want to specifically note that “I’ve assumed the initiative to decrease time in training was not a significant impact to these results”.
If you call them out before someone else does, you will retain your credibility longer.
Got it now?
Three steps, not so bad, right?
Here is the summary of takeaways:
- Change your mindset — look for a number whenever you hear the words improve, change, or reduce. If you are making a change, there’s a way to quantify it.
- Do the math — you don’t need calculus, just look at the numbers from x to y in z time. And maybe learn a percentage or two. And definitely learn how to ask your favorite AI tool for help!
- Evaluate the answers — create a story of where you started, where you ended up, and how you got there. Show the pictures of the data that help support the story. Try a few different versions to make sure the picture shows your story.
Tell your story and illustrate it with the data summaries. Your boss, colleagues, and clients will appreciate them.
Congratulations. You are now a data-driven manager.
I’d like to hear your struggles in the comments. ⬇️ Is there a change happening and you don’t know how to quantify it? Is there another simple piece of math that wows your team? What else have you started that makes you look like a data-driven doer?
Thanks for reading. If this is helpful or interesting to you, I’d appreciate your feedback. If you want to improve your company’s data usage or need help from other leadership/management, please contact me at [email protected].
