Turn Data into Action: An Application of McKinsey’s “Insights Factory” Approach

Sales and marketing organizations have an abundance of data these days, and while a lot of effort tends to go into managing it, its true purpose – and our true challenge – is using it to create value.

The goal sounds simple on its face: transform your data into insights and then into action in a fast, repeatable manner. Following the “insights factory” approach first described by McKinsey, I’ve found that true value can be achieved systematically in four steps. After successfully experimenting with this approach, I’ve decided to share its practical application in sales and marketing.

I have been working closely with data in the marketing and sales organization of an enterprise SaaS company for four years now. I started with deep dives into website performance and marketing funnel data, then eventually dove into more exciting data about the sales pipeline and marketing attribution. I’ve worked with data from Google Analytics, Marketo, Dynamics CRM, and other tools, feeding all that information into Power BI, Microsoft’s interactive data visualization tool.

Long story short, I’ve seen both the ugly side and the beautiful side of data – and no matter how great your data, it’s useless if it isn’t getting turned into action.

Here are the four steps to working with data successfully, with examples from my own experience.

Step 1: Decide What You Want to Produce

The best way to start is figuring out where you want to end up.

Establish a clear understanding of what you want to achieve, what questions need answering, and what actions those answers will drive. McKinsey advises to “prioritize questions that address the largest economic opportunities and lead to practical actions.” Once we know what we’re looking for, we can set up our “insights factory” to zoom in on those specific questions.

We can also approach this by formulating hypotheses. When we already have significant knowledge on a particular subject, we can create a provisional explanation as to why something is happening in our process. Then we can collect and analyze data to find out whether it supports or negates the hypothesis.

Example: Which acquisition programs are most effective?

My company’s marketing department wanted to make data-driven decisions about where to make additional investments and where to pull back. If you tell me that you want to know how your programs are performing, I’ll respond with a series of questions. To determine the right set of metrics, I need to know what you plan to do with that information and what actions you intend to take.

Allocating budgets to the most effective acquisition programs requires us to know the cost and make-up of the leads we are acquiring into our database. To determine that, we selected metrics like cost per acquired lead; percentage of acquired leads from our Total Addressable Market (TAM); cost per acquired TAM lead; and lead breakdown by job level, function, region, and industry.

Our insights are intended not only to guide marketing decisions, but also to ensure that we have the right data in place. Which brings us to Step 2.

Step 2: Source the Raw Materials

I’ve often seen projects start with “let’s collect this data” or “let’s create these fields” or “use all these metrics from Google Analytics.” We’re constantly bombarded with information, so it pays to remember that power is knowing what to ignore.

Once Step 1 is done and we have clear goals, it’s much easier to gather our data. We’ll also be less susceptible to information bias – that is, gathering more information than is necessary to inform decision-making. It’s best to start with a selective dataset – get something good enough, fast, so we maintain the ability to act quickly. We’ll add layers of good data as they become available.

If data is missing, we should consider how long it will take to collect enough of it. Then we’ll prepare the data to make sure it’s useful, clean, validated, and consistent.

Example: What data is available about acquisition programs?

Data for our acquisition report is sourced from Dynamics CRM and is built through integration with Marketo and Google Analytics. We set up the data structure long ago, enabling us to track the channels that generate leads – chiefly, advertising, content syndication, conferences, and webinars hosted by our partners.

As for programs, the data structure isn’t quite there; it’s too granular and inconsistent, making it unreliable. So, we went with what was available and reliable – data identifying the most effective channels, not programs.

Step 3: Produce Insights with Speed

Now we want to produce insights quickly so that we can make decisions and execute. Personally, I’ve struggled with this step due to my perfectionism – which is actually fear of failure.

The guidance I follow is again McKinsey’s: act like a start-up. In the words of Facebook, “move fast and break things.” Prioritize speed and don’t let the perfect be the enemy of the good. The pursuit of perfection – particularly right from the beginning – can be paralyzing; instead, form what McKinsey calls a “test-and-learn culture” – acting rapidly, learning from your mistakes, and adjusting accordingly as you go along.

This is a critical lesson. Failure while learning is acceptable, and action is empowering. So, test and learn – using the “good enough” information available now to inform specific actions.

Example: How did we generate rapid insights about acquisition?

Our first insights were about data quality and integrity, so our immediate action was to redefine our data collection and management processes to gain, for example, consistent industry data.

In two days, we had metrics calculated only for those leads linked to our DCRM accounts at that time – about 33% of all generated – but we had no metric for cost by channel yet. Over the next two weeks, with changes to processes and data, we generated good enough information to feed all the metrics. That gave the marketing department a rough idea of channel performance.

Step 4: Deliver the Goods and Act

With Step 1 done well, it’s going to be easy to come full circle with this final step. After all, if we start with a clear understanding of what our “clients” in sales and marketing need, we can provide insights that drive action.

McKinsey explains that for the “insights factory” to perform optimally, the sales and marketing teams require easy and timely access to its results, which typically means you need some interactive tools. But you can’t just push tools on your company or your customers; rather, your duty is to identify their needs and fulfill them. If you fail to do that, you’ll be at risk of providing analysis that’s seen as impractical, unclear, or irrelevant – or maybe all three – which means it’ll be ignored.

Example: How did we deliver our acquisition insights?

Our interactive tool for delivering insights across teams is Power BI, but in this particular case, we delivered the report in Excel. Its insights enabled marketing to pull funds away from low-performing channels and redistribute them to high-performing ones – those that generate leads with high job levels in our TAM, and at low cost.

Surprisingly, LinkedIn was giving us lower job levels at a higher cost per lead, so one immediate action was a reduction of our spending there. Meanwhile, we were getting high performance out of our content syndication channels, so we’ve increased our focus on them.

Follow these four steps in your quest for insights, and you’ll be off and running. The first step is the most crucial of all, but each one is vital to maximizing the value you’re able to offer. Now get out there and turn your data into action!