top of page
  • Writer's pictureCsaba Tamas

A Zero to Hero Analytics Framework to scale your business

Updated: Jul 19, 2021

Becoming a truly data-driven company is the “superfood” of customer-driven business growth. According to a MicroStrategy study, data-driven decision-making leads to improved efficiency and productivity, better financial performance, improved customer experience, acquisition, and retention. At the beginning of my analytics journey, I wasn’t sure where to start to get the most out of data. I read several books and tens of articles. I synthesized my research into a practical, Zero to Hero step-by-step blueprint. Read on to gain an insight on the “Scaling Analytics Framework” that can fuel your engine of growth for years to come.

Read the prequel to have the "HOW" covered too.

If you think about it for a moment, building and scaling a company often follows a similar path to the one of developing and launching a rocket. It all starts with a bold idea, a series of iterative design and test cycles. Many of these end with a failure and the hope that tomorrow will be better. We expect success will knock on the door before we run out of money. Most product launches are modest events at first, but successful ones are slowly building momentum. Day after day, one step at a time. It is a continuous climb, much like the rocket that leaves behind the pulling gravity of Earth. Just like the rocket follows a well-calculated trajectory, entrepreneurs are also defining a path towards the envisioned destination and are gazing at the climb on a central dashboard. But on a second look, scaling a company might be even more challenging in some respects: it is more like launching a rocket to the Moon and trying to figure out most of the space-flight science during the flight.

Analytics plays a pivotal role in this process, ensuring that your engine of growth has enough torque while not blowing up halfway to the exit event. First, analytics helps you to identify and describe the forces that influence your engine of growth and your crew. Next, it helps you to calculate the shortest path to the exit event. It provides a framework to identify and measure the activities that will bring you closer to the goal, one step at a time. And finally, analytics helps you to track progress along the way. It shows if the subsystems of your business are performing under nominal conditions. Analytics provides early alerts: be it too much fuel or oxygen consumption, or approaching the surface of the Moon too fast or too slow. In the case of the business: it provides early alerts on change in the revenue trends or customer engagement.

The existence of a real-time, pervasive analytics framework, is paramount for delegating high-velocity decisions too. It also ensures conceptual integrity and alignment among teams. The cause and effect of hundreds of variables suddenly become transparent to everybody. Decisions become pragmatic and thoughtful, beyond bias and gut feeling.

And that is the purpose of the “Scaling Analytics Framework” that has 5 major steps: 1.) exploratory analysis — understand what your customer wants; 2.) define your objectives, the precise direction where you are going — we call these output metrics 3.) identify the performance indicators that lead you to your objectives — we call these input metrics 4.) create situational awareness in your organization using a cause-effect network of key performance indicators; and finally in step 5.) track your performance and the validity of you your metrics and course correct when needed;

This system provides you instant feedback on course correction beyond the control room. Thus, individual teams will better understand the impact of their decisions.

Step #1 - Exploratory analysis: figuring out what customers want

Some business failure is a result of building products and services that nobody wants (while the entrepreneur is convinced about the viability of the idea). That’s why I encourage you to stop for a moment and distance yourself from your idea. Before you jump into the joy of building, try to figure out what your customers want. Or even better: what they need (a small, but important nuance).

Make sure you make what you can sell and not try to sell what you can make. And that means figuring out what people want to buy. Qualitative customer research is more important at the beginning of the product life cycle. My favorite book on the topic is called the Mom Test by Robert Fitzpatrick.

As soon as users start interacting with your product, you can shift to quantitative research using the ever-increasing amount of data. Now, you can run an exploratory analysis and look for a correlation between dependent metrics and desired outcomes. Some questions you could ask: what marketing activity shows the closest correlation to revenue growth? Which communication channel brings more customers? What product-feature increases engagement?

When you go beyond correlation and find causation, you know you found a Jackpot. If you can identify what drives the behaviour of your customers, you can predict and influence the future.

At this stage, however, we are targeting a potential correlation between your present actions and future results. In other words what input metrics lead to what output metrics. This is what we cover next.

Step #2 - Define desired outcomes, the output metrics (a.k.a. lagging indicators)

Let’s start with the desired business outcomes, like revenue and growth-rate. When you observe these, they are the result (output) of your past activities. That’s why we also call these “lagging indicators”. They show a reality that lags behind actions delivered in the past.

The series of output metrics are representing your path to the Moon. The destination of your rocket. Here you answer WHAT would you like to achieve, as opposed to HOW can you achieve it.

You can start by asking yourself: what is the business outcome you want to see at the end of this period (quarter or fiscal year)? The answer can be anything in the range of the number of activations, the number of monthly active users, retention rate, net promoter score, app store rating, or revenue growth.

I recommend defining at least one output metric in all key areas, such as sales, marketing, product, and operations. Some standard metrics can get you a jumpstart, be it the so-called “Pirate Metrics” or the “HEART Framework”. The Pirate Metrics include Activation, Retention, Referral, and Revenue. These are focused on your sales performance. The “HEART Framework” on the other side is focused on your product performance. The metrics include Happiness (attitudes or satisfaction), Engagement (or user involvement), Adoption (number of new people using your product or feature), Retention (users who stick with your product), and finally Task Success, which covers behavioural metrics of UX such as effectiveness, efficiency, error rate.

Output metrics are usually easy to measure. On the downside, it is hard to identify the actions that can improve them. Think of revenue: it is simple to measure, but there’s no single conclusive step that you can follow to improve it. Since we are looking for a system that can give us direction and magnitude, we need to keep going.

Step #3 - Identify input metrics that influence future outputs (a.k.a. Leading indicators)

If output metrics — discussed in the previous section — are about defining the WHAT, input metrics are all about the HOW. There are at least two reasons why these are the most important metrics: a.) They give early indications of performance, thus they can predict the future, and b.) opposed to the output metrics, these are actionable. They describe the actions you can take to improve future outcomes.

Many businesses stop at defining and tracking output metrics in the boardroom. That’s not great, because output metrics can only be observed after the fact and you need input metrics that you are in full control of.

Take the example of the Net Promoter Score (NPS), which captures the customer’s sentiment towards your product or service. We can only observe its value at a time when the underlying customer sentiment is already settled. Input metrics — on the other hand — like “Average First Call Resolution” or “Daily Active Users” are early indicators that predict future NPS values.

But how to find the right input metrics for future NPS? Just ask the magic questions: What are the customer behaviours that predict satisfaction? What are performance metrics that will have a positive impact on customer satisfaction?

You could look at customer-service quality for example. Ask how promptly your teams are responding to a complaint? Is there an increase in the volume of outstanding complaints? What are those complaints about? Is there a correlation between the increase of complaints about a particular product and a decreased NPS for the very same product? What is the number of average distress and inconvenience payments? What is the elasticity between the inconvenience payment and the NPS score? What are the root causes of those complaints? What is the value of successful resolution for those complaints?

Note: leading indicators are a.) early predictors of future outcomes and b.) measure activities that you are in full control of.

Step #4 Create situational awareness by ensuring business observability

“Entrepreneurs are particularly good at lying to themselves. Lying may even be a prerequisite for succeeding as an entrepreneur — after all, you need to convince others that something is true in the absence of hard evidence. […] As an entrepreneur, you need to live in a semi-delusional state just to survive the inevitable rollercoaster ride of running your startup. Small lies are essential. […] But if you start believing your hype, you won’t survive. You’ll go too far into the bubble you’ve created, and you won’t come out until you hit the wall — hard — and that bubble bursts.” 
 Croll and Yoskowitz / Lean Analytics.

Yes, you need to deceive yourself a bit, but not to the point where you’re jeopardizing your business. That’s where continuous, real-time situational awareness comes into the picture.

Let’s package your business intelligence into a graph of input and output metrics. I call this a Key Performance Indicator Tree (KPI-Tree in short). Combining the business outcomes (the output metrics), and the input metrics (the actions you need to take), we build a funnel or better said a tree structure of KPIs that are representing the trajectory of your scaling business. A KPI Tree is a powerful visualization tool that helps to break down the organization’s objectives into granular activities. This is somewhat similar in concept to the Objectives and Key Results (OKR) used by many startups today. It differs from it by providing higher granularity, immediate visibility, and tracking continuous performance metrics (not just new growth areas).

These are the steps to build a KPI tree of your own:

  1. The root of the tree is the output metric, the business outcome;

  2. The leaf nodes of the tree are other output metrics that are part of the main output or the input metrics that lead to the desired outcome.

Let’s use an example to illustrate the concept: say your goal is to close 50 new deals (a lagging indicator). With a six-month-long sales cycle, you would only know if you succeeded at the end of the 2nd quarter. To build an early warning system that gives you time to steer the rocket, you could set leading indicators to track your progress. Think about it: what could be a good candidate for the leading indicators in this scenario? How about Pipeline Volume or Number of Calls/Meetings/Emails per sales rep. By tracking these metrics, you build a real-time perception of progress. You will know ahead of the period if you need to increase outreach efforts to meet your Deals Closed goal.

The combination of multiple KPI-Trees will lead to high business observability, which is a prerequisite tool for high-velocity decision-making.

The web of KPIs interwoven into the fabric of the organization leads to increased business observability and this, in turn, leads to an increased ability of the organization to better understand itself and adapt fast to changing market conditions.

Step #5 Measure, Refine and Improve

At the beginning of your product design process, you will make many hypotheses. You will try to define the customer needs and identify potential input metrics that lead to your desired outcome. As the business evolves, you will need to revisit your output metrics and redefine your KPI-Trees as you validate many of your early hypotheses.

Good metrics that help you grow fast are actionable, comparative, reliable, understandable and usually, take the form of a ratio.

Once implemented and maintained, the “Scaling Analytics Framework” will provide purpose and direction for your teams. It will assure a holistic understanding of the value streams and lead to real-time business observability as a basis of high-velocity decision making. These elements are the key ingredients of an autonomous organization, which scales faster due to its ability to react to changes in the environment autonomously. And that is the basis of a data-driven organization and a key capability that will contribute to your sustained growth.

This article was first published on the Entrepreneur's Handbook on Medium.


bottom of page