In this article I will review a methodology by which you can realize a massive return on your investment in data & analytics.  Most organizations that we encounter are not realizing most of the value of their data before they engage with us.  We help them not simply implement technology, but doing it with best practices and design patterns that come with decades of combined experience.  We wrap that with an overarching methodology that I put together after a couple decades of building and leading data teams at some innovative companies like Salesforce, PopSugar, and now of course Data Clymer.  Whether you are just starting out on your data journey, or you’ve been doing data for years, making sure you follow this process is really going to pay off.  

When thinking about how to be successful with data, most people go straight to the discussion about technology and solution architecture.  The questions we typically get here at Data Clymer are focused around  best-in-class tools for a modern cloud data stack.  And of course we walk our clients through the pros and cons of world class solutions like Matilion or Fivetran and DBT to ingest and model data, and why a cloud data warehouse like Snowflake is an amazing choice, and how business intelligence tools like Sigma or Looker are wonderful for getting data in front of end users self-service.  While these are all important discussions, what’s even more important is putting together a comprehensive data strategy around the technology. 

Which brings me to what I call the “Actionable Analytics Cycle”.  This methodology is all about people and process.  If you do all of these steps well,  you’ll start seeing that return on investment.  Now, there are a host of best practices and design patterns that should be followed at every step of the cycle.  I’ll briefly mention some of them today, but I’m gonna reserve a deeper dive for some future articles.

The goal of this process is to take actions based on your data insights. Without action, you’re not going to see any value from your data.  So where do we start?

Questions:

Well, you start by asking questions.  By this I mean, what are your organization’s goals?   Whatever your goal-setting methodology, be it OKRs (stands for objectives and key results) or something else, you need to identify the most important objectives so we can then devise a strategy to apply data. 

Metrics:

Once you’ve asked the right questions; we need to identify some metrics that will measure success and uncover drivers of that success.  I find that teams often don’t spend enough time thinking through appropriate metrics.  For instance, during the eight years that I spent at Salesforce building the data team, I would routinely work with product executives who often suggested a vanity metric like the growth in user licenses over time.  But that wouldn’t tell them much about the success of their specific product within the Salesforce suite, so it wasn’t actionable.  Something like % of users regularly using a key feature would be much better.

One of the best ways to know if you’ve identified a good metric is to ask yourself, “If that metric were measured today and I had the results in front of me, what would I do about it?”   If that’s not very clear, then your metric is probably not actionable.

Trust:

Okay, so you have some solid metrics, you might think the next step is to go get your data.   But you can’t ignore this next step:  trust.  It’s imperative that you ensure the data is trustworthy first.  You see, lack of trust is the #1 killer of data initiatives.  So, first evaluate the quality of your data sources, then think through ways to ensure accurate data processing, which includes checking for completeness.  Design robust, self-healing data jobs.   Design and build a system health dashboards.  Think about how you can ensure the data teams will be the first to be alerted about any issues?   All of these things should be included in your project plan.

Data

You’re finally ready to go get the data!  So, this step is all about building a trusted data pipeline that ingest raw data from your source systems into your data warehouse, where all data needs to be centralized.  I can’t emphasize enough that centralization is key.  You need to be able to join data together in a single SQL statement to rapidly analyze it.   Otherwise the time it takes to complete an analysis goes through the roof.   

Modeling

Raw data nearly ALWAYS has to be modeled.   What does data modeling mean?  It’s the process of transforming a lot of complex raw data tables into a set of much more analytics-friendly, easy-to-use tables.  Data modeling is what I often call the “art” of what we do at Data Clymer.  It can be the most complex part of an implementation because raw data is often not structured well for rapid analytics.   There are some exciting new modeling patterns that have become possible thanks to the raw power of modern cloud data warehouses.  More on that in another article.

Analytics

So, once you have your data centralized in a modern cloud data warehouse and it’s modeled to expedite analytics, it’s time to start measuring your metrics and analyzing the data to get some answers to your original questions.   This is the fun part, right?  There are a vast number of approaches and techniques that can be employed to analyze your data, and it usually takes a very different kind of skillset to do this well vs. what it took for the data engineering work in the previous steps.  This is why is usually take an entire team of people to do data right.

I have to mention here the concept of data democratization.  It is imperative that part of your strategy is to get data into the hands of as many people as possible.  That way, they can answer most of their own questions self service, alleviating analysts to tackle the harder problems.  Plus, democratizing data allows people to interject data into every aspect of their job.  This is how you create a data culture in your organization.

Answers

So now you’ve turned your data into information, and that information into answers.  

Actions

That’s not the end of the story.   The question is, what are you going to do about it?  How are you going to leverage the insights to fix a problem or improve a situation? Will you alter the way you market and sell to customers, or what services you provide, or how you provide them?   Will you build new products, or change internal processes?   Taking an action is the only way to get any value out of this entire process.    And when you do, be sure to measure results.   And then, keep asking more questions.  The cycle never ends.  


Aron

 

About Aron

An executive leader passionate about extracting maximum value from data, Aron founded Data Clymer to help organizations implement optimal data strategies and instill a data-driven culture. With over two decades of experience, he more recently spent seven years building and leading the Product Intelligence team at Salesforce, and another two years doing the same at PopSugar.


Data Clymer is a premier consulting firm specializing in full-stack analytics and data culture transformation. Our proven methodology includes full data stack implementation, data democratization, custom training, and analytics to enable data-driven decisions across the organization. We have curated a set of best practices from our deep expertise in Snowflake, Sigma, Looker, Matillon, DBT, Fivetran, Tableau, BigQuery, and more. 

If you need professional help, contact us or follow us on LinkedIn.