Join top executives in San Francisco on July 11-12, to hear how leaders are integrating and optimizing AI investments for success. Learn More
According to a Gartner study, about 40% of enterprise data is either inaccurate, incomplete, or unavailable, which results in businesses failing to achieve their data-driven goals.
As a head of analytics, do you doubt the worth of your expensive data investment in regards to accelerating the growth of your organization? Are your data science projects taking too long to implement and also having minimum impact? Do you have numerous pending Jira tickets that you are unable to prioritize? If you answered yes to these questions, you are not alone. In fact, most startup business leaders are almost certainly sailing in the same boat.
Have you ever wondered what the other business leaders are doing differently?
The key to analytics success is learning how to have a business approach to picking the right projects that will drive maximum impact among the thousands of analytical tasks that might come up during the year. In a world of fierce competition, especially for startups, a goal-driven approach is often the only difference between success and failure.
Join us in San Francisco on July 11-12, where top executives will share how they have integrated and optimized AI investments for success and avoided common pitfalls.
4 tips to get analytics ready
As a business, you may have enormous amounts of data flowing in. Ideally, this data should be leveraged to derive valuable business insights to drive growth. But there is a huge difference between expectation and reality.
To start with, you probably have a number of questions that you can ask about your data. Out of these questions, some may drive growth, while some do not fetch any value. Some can be answered by building thoughtful dashboards, while some require deep data science to get to the answers. But data in its varied and complex forms can overwhelm analysts and organizations. As a result, organizations may lose focus and waste valuable time answering all the questions. In such a situation, it is only logical that you prioritize the questions based on the estimated impact.
Here are the four most meaningful steps that can give you maximum return from your data investment.
Tip 1: Invest in technology consolidation and data architecture
Let’s assume that you are the head of analytics of a hypothetical food delivery app — FoodNow.
Restaurants took a big hit during the peak of the pandemic because people were reluctant to go out. Only takeaways were working. This is when you launched the FoodNow app, which connects restaurants with customers.
Your business started flourishing as you used Google ads to reach more customers. Your customer base has been growing. As you grew, you started storing your data in AWS S3 and Snowflake, managing your new customers through Salesforce, and started using analytics tools like Amplitude and Google Analytics to collect, store and analyze the data coming from different sources. You used that data to track the average delivery time, your growth in various places, and the restaurants that grew along with you.
You are now at 10,000 customers and you are using basic data and tracking and everything is going great. But are you going to use the same strategy to grow from 10,000 to a million customers? Likely not. You now need to start understanding who your customers are.
Seeing first-level trends in the amplitude dashboard is not going to suffice to understand your customers and you may need to do more advanced analysis like customer segmentation. At some point, the founders are going to come to the head of engineering or IT and ask how we can use this data to understand our customers and their needs better? Like any executive team, they would want insights fast as well. However, you are collecting millions of data points a day. How can you turn that into meaningful insights, fast?
As you grew, your tech stack grew, solving for each use case, as it arose. That resulted in disparate data sources and multiple data definitions. The growing data volume created inefficiencies in storage and computation and resulted in long cycles of project delivery and slow dashboards.
If you are at this stage, it’s time to invest in data and technology consolidation, and develop business-in-data architecture. It’s time to evaluate cloud technologies like GCP, AWS, and Azure to bring all your data and processing into a single system. However, unless you do proper data architecture design, your systems will be extremely slow and inefficient. So it’s important to assess and design proper data models and think about governance, security, and other MDM aspects so you can build to scale.
The other key aspect to remember as you go through technology investment and architecture is that not all data has the same value. Some (top 300-500) metrics are going to be used daily and thus need to be stored in a well-designed and fast database vs. other data points that can sit in a data lake.
In the next tip, we will talk about how you can use a business-in perspective to identify the top metrics and the top projects which move those metrics — a.k.a. analytics agenda. Then we will discuss how to use the analytics agenda to create a Single Source of Truth (SSOT).
Tip 2: Create an analytics agenda
You now have an overwhelming amount of data at hand which can come in handy to answer an unending list of questions. The questions could be, for example, how do I improve my delivery time? How do I track the current location of the delivery? Which route is the shortest? It makes sense to prioritize your questions based on the estimated impact. And to achieve this, the analytics agenda comes to our rescue.
There are 3 key steps in the analytics agenda:
- Understanding the KPIs
- Identifying the driver metrics
- Figure out the projects which line up with the KPIs
For our use case of the food delivery app, the two most important KPIs would be:
- Revenue growth
Let’s take the first KPI.
The L1 driver metrics for the revenue growth would be:
- Number of customers
- Orders per customers
- AOS (Average Order Size)
- Revenue per AOS
But what drives these metrics?
The L2 driver metrics for the number of customers would be:
- New customer
- Returning customer
Then ‘New Customers’ you acquire would be a function of the Acquisition campaign, i.e., total eyeballs, clicks, landing page, and final order placement. And so on… Building your metric driver hierarchy (L1 to say L10) will create your measurement framework.
Now the final step, figuring out the projects that line up with the KPIs.
Now you can use the measurement framework to identify analytics projects that drive key driver metrics and estimate the value of the project on the top KPI. For example, if there is a project whose goal is to potentially increase the new customer acquisition rate by 1%, and that 1% equals $10M, the estimated impact of the project is $10M. Whereas another project may focus on increasing orders from returning customers by 2% which is estimated to drive $8M in incremental revenue, then the estimated value of that project is $8M.
Similarly, you need to calculate the estimated impact of all the projects on hand and figure out the top 10-15 projects that line up with your KPIs. This gives you a fair idea of your analytics agenda. Once you have figured out your agenda, you can easily estimate the amount of incremental value that you could drive at the end of the year.
Now that you know your analytics agenda, you can use a hypothesis-driven framework like BADIR to identify the critical metrics needed to solve those use cases that you are going to learn further in this blog. Those top critical metrics then feed into your SSOT which you systematically store in your database whereas the rest of the data can sit in your data lake.
Now that you know how to draft our analytics agenda and have a line of sight to SSOT (easy access to data), the next question that may pop up might be, “Are my analysts ready to deliver on the analytics agenda?” which brings us to the next important tip in the process.
Tip 3: Empower analysts to ask the why behind the what
Managers today have to do more with less, and get better results from limited resources, more than ever before — Brian Tracy
Managers need to take action to satisfy their organization’s mission and vision. This can be challenging, especially in the field of analytics where resources are limited and critical resource strategies need to be reinforced to accomplish the tasks in time.
Let me explain.
For a moment, let’s assume you are the head of analytics for FoodNow.
Before taking this discussion further, let me ask you a quick question: what do you think is your role as the head of analytics? If you are like most people, you might be thinking that your role is to support marketing objectives or help product departments.
Now, here’s a surprise for you: though supporting the goals of product heads and marketing heads is part of your responsibility, we believe that your primary role is to monetize the data you have at hand. In this case, you cannot afford to lose sight of the 15-20 major projects you have identified as part of your analytics agenda.
Now let’s say you have agreed that your role is to monetize the data. The next big question would be “How do I operationalize the agenda?” How can you constantly work on these strategic projects while being bombarded by day-to-day tactical questions from the marketing and product departments?
Here are the tips to operationalize your analytics agenda:
- First, you want your team to quantify the expected impact or outcome of the projects before they commit to them.
- Analysts should be able to prioritize impactful projects.
- You might need to permit your analysts to say no to projects that don’t drive impact.
- Ask your analysts to be involved in the projects from the stage of defining the problem itself and train them to not just work on the ask but to understand the intent behind the ask and redefine the problem statement if required, so that they are answering the right business problem. Understanding the why behind what helps the analyst to work on the real business question rather than the initial ask.
- Make sure that your analysts are spending at least 40% of their time on one of the top strategic projects from the analytics agenda that drive maximum impact and the other 60% catering to day-to-day tactical support.
By creating this discipline, we are ensuring that our analysts are continuously growing and evolving as a high-performing team.
Even if your analysts are able to ask the why behind the what and prioritize analytical tasks, how do you ensure scalability? Do you have a standard process in place? This brings us to the next tip.
Tip 4: The analytics process
Chaos is the enemy of growth and process is an ally.
The business world feels like a rat race. Everyone seems to want answers yesterday, they have tight timelines today and the data is questionable. Without a process for analytics, a lot of projects will get done with zero value in increasing the ROI.
Let’s zoom in on that.
Going back to the FoodNow example: let’s say that the head of finance is asking for an LTV model to evaluate the investment in acquisition over the last year. So, the analyst goes and looks at acquisition and various retention curves and comes back with the answers. The head of finance puts these numbers in an Excel sheet and then says the retention curve you are using might not be the best one since we changed our pricing model last year.
After a few weeks of back and forth, the head of finance shows the numbers to the head of product who does not agree with the methodology. He says, there’s a problem, and your retention model is no longer holding. Unexpectedly, at the same time, the marketing person shares about going viral on TikTok during a certain period, and the retention numbers or the numbers we are assuming are not reflective of what’s realistic. Imagine how chaotic the situation would become and months would pass by without the project getting materialized.
Amid all this chaos, you pull aside an analyst from a random organization and ask whether he is enjoying his job. The most definite answer will be a “no.” He may say that the process keeps changing so often that he has no idea what he would be working on tomorrow.
Now, enquire a stakeholder from the same organization whether he is happy that an analyst is working on his project, say, an LTV model, I am sure the answer will be a “no” again.
He may be thinking that the analyst is not skilled enough to bring thought leadership to the table. The same would be the case with the engineers, product heads, and the organization’s directors. This is commonplace. What is more common is cluelessness among various departments of the same organization about the stage and purpose of the analytics project. Each thinks that the other is not contributing enough. There is neither a process nor alignment here. This is so many analytics projects break down at this stage.
Using the BADIR framework
To avoid this scenario, I want to bring to your notice one of the most robust analytics processes which we mentioned before. The methodology is BADIR, an acronym for five steps (B: the business question, A: the analysis plan, D: the data collection, I: stands for deriving insights, and R: is for making recommendations). There are various subsets within each stage. This process is also discussed in detail in my book Behind every good decision. Chapter 4 in the book talks about the entire BADIR framework and many methodologies and their usage.
In this process, everyone, including the managers, engineers, stakeholders, and product/marketing heads on the other side will be roped in for project discussion from day 1. The analyst will start by asking questions about the need for this analysis? Who are the stakeholders? What actions do they want to take? This helps the analyst draw all the information required to frame the right business question. All of this information makes the goal clearer. It’s like working backward and understanding what it is that you want to get at the end of this project?
In the analysis plan, they draft the analysis goal, the hypothesis (drivers of the chosen KPI), the methodology, risks and constraints, and the timelines. It is extremely important to take this step very seriously because a well-drafted analysis plan is a key to the success of the project. I suggest that the analyst should not start working on the analysis until the analysis plan is well-drafted, understood by all stakeholders, agreed upon, and signed off by stakeholders.
Data collection and analytics
Collecting the right data based on the analysis plan, from the correct data source and then validating is an important process. Remember, getting good results and actionable insights do not depend upon choosing the correct model and tuning it for accuracy, it first depends upon the quality and accuracy of the data that we input into the model.
Once the analysts have cleaned the data, now they can use defined recipes on the agreed methodologies from the analysis plan step, using all hypotheses, to systematically arrive at insights and quantify the impact of the insights.
Finally, they make an actionable recommendation. Along with the recommendations, it is important to quantify the impact of each recommendation so that we are ensuring that their recommendations are having a direct impact on the business.
This analytics process brings order to chaos aligning stakeholders every step of the way with clear objectives and processes so data can drive impact. This is why a process is crucial for successful analytics.
Like a bodily dysfunction, failure warns us that something has gone wrong within a system. The key is to identify the cause and take needed action. Success in startups and sustainable development can only be achieved with smartness, observation, and persistence. Asking the right questions, easy access to data, empowered analysts, and a solid analytics process are the keys to unlocking the power of your data.
Piyanka Jain is president & CEO of Aryng. Sahaj Harnal is a data scientist and is currently working with Aryng as a consultant – data science. Swarnim Shrey is an analyst and customer success manager at Aryng.
Welcome to the VentureBeat community!
DataDecisionMakers is where experts, including the technical people doing data work, can share data-related insights and innovation.
If you want to read about cutting-edge ideas and up-to-date information, best practices, and the future of data and data tech, join us at DataDecisionMakers.
You might even consider contributing an article of your own!