How to Turn Mounds of Data into Usable, Meaningful Insights (2022 Guide)

Uwemedimo Usa
February 24, 2021 ·
How to Turn Mounds of Data into Usable, Meaningful Insights (2021 Guide)

Are you overwhelmed by all the data in front of you?

There’s an unprecedented richness of data out there, but how can you turn it into actionable insights?

You see, the journey from data to insights is littered with challenges. You will need a potent set of steps to combat them.

Here we’ll show you how to extract insights from mounds of data, cut through useless information, and make faster data-driven decisions in your organization.

What Are Data and Insights?

Before we get into it, let’s define what we mean by data and insights.

  • Data are facts and statistics collected through observation. They could be numbers, text, images, audio, etc.

    Let’s put this into perspective:

    Supposing you own an eCommerce store and have Google Analytics (GA) activated, each user that navigates your online store leaves some digital footprints that are picked up by GA.

    This covers demographic data, devices, browsers, etc. If you see these raw pieces of facts, it would probably look gibberish as they’re sometimes unstructured and without context.
    Raw data Google Analytics insights
    Source: World Bank
  • Information is a product of refining this data and giving it structure and context. This way, data makes a little more sense to the naked eye.

    In the case of your eCommerce store, an example would be your GA dashboard. When all the data is put together and presented in context, it is consumable and you can draw insights from it.
    Information Google Analytics data
    Information is data you can read
  • Insight is valuable knowledge gained from an understanding of information. When you consume information (or data) and accurately interpret it within its context and other information available, you arrive at insight.

    In the business world, insights are the entire point of collecting data. Think of insights as looking into the inner workings of the operation you’re observing. They tell a meaningful story from data.

Example of an Insight

Identifying through effective data analysis that 97% of your customers find you when they’re planning their wedding is an example of insight.

On its own, this is merely information that’s cool to know.

But if this is used to craft a plan that brings tangible value to your brand and customers, that’s actionable insight. For instance, you could use this information to plan an ad campaign targeted at people about to tie the knot.

You already know this audience is drawn to you; thus a high ROAS (Return on ad spend) wouldn’t be too surprising.

Before data is transformed into insights like this, they have to first be collected and then analyzed.

Various Methods in Which Data Can Be Collected

Most businesses collect data from multiple sources using different methods. And each method comes with its own set of rules.

Google Analytics, for instance, gathers data using JavaScript and a tracking code.

Collecting data Google Analytics insights

Once the tracking code is added to your page, the script sends data to Google servers. This includes page data (URL, title), browser data (viewport, screen resolution), user data (location, language), and many more.

It’s similar to social media analytics on Facebook, Twitter, Instagram, and other sites. You can also collect data from:

  • surveys
  • market growth statistics
  • transactional data tracking
  • customer feedback analysis
  • subscription and registration data, etc.
Data sources insights

From here, data moves to the next step.

Data Analytics vs Data Analysis

Contrary to popular usage, these terms don’t mean the same thing.

  • Data analytics is the science of collecting and using data. It is everything between collecting raw data and taking action from it. This includes the collection, organization, storage, and analysis of data using machine learning, statistics, and computer-based models.
  • Data analysis is a subcomponent of data analytics. Data analysis is a process of examining, cleaning, transforming, and organizing data with the end goal of extracting valuable information and using it to inform decisions and act.
Data analysis vs data analytics

In a typical organization, data scientists, executives, and managers are usually the only ones using data analysis to derive insights.

An efficient data-driven organization should distribute access to data and understanding of data to all of its members.

Learn what data analysis is and how it differs from data analytics in this video from Convert’s A/B testing course.

This brings us to a game-changing concept: data democratization.

What Is Data Democratization?

Data democratization means making data available to everyone within the organization without the barriers of ultra-specialized expertise. This means anyone in your organization can access data, understand it, and use it to make decisions and recommendations.

The idea is that the more hands-on-deck (data-wise), the faster the company adopts a data-driven decision-making culture from top to bottom.

But there’s a catch.

With this level of access, it becomes tougher to maintain data security and integrity. There’s also the possibility of data being interpreted wrongly by someone without the expertise of a trained analyst.

Still, data democratization is a key driver in making smarter and faster data-led decisions and improving customer experiences.

The marketers at the Royal Bank of Scotland showed how efficient it could be to involve their non-marketing colleagues in the customer experience optimization process.

Challenges in Generating Insights from Data

The path from data to insights is riddled with challenges. So much so that alternatives to data-driven actions seem more attractive.

Marketers, data scientists, business executives, and other professionals who work with data on the daily seem to agree.

I ran a quick poll (here and here) that revealed data validation is the biggest challenge for 39% of them. Only 11% attributed their difficulties to the volume of data. 28% went with integrating data from various sources, while 22% cite the time and effort involved.

Challenges in generating insights from data

Apart from these four, other challenges with transforming data to actionable outcomes include:

  • Inaccessibility of data
  • Poor quality of data, and
  • The pressure to deliver ROI

For Steven Alexander Young, Founder of Challenger Digital, the biggest challenge is isolating the variable behind a change in performance. Analytics data do not always tell the full story:

When traffic dropped here, was it because someone made changes to the page (and if so, what)? If the page didn’t change, did a competitor happen to beef up SEO and overtake you (and if so, who)? (…) Even when I can get a client on the phone to provide details and rule things out, they often have to go on their own goose chase within their team to track down answers. Of course, parallel to this is the ever-present possibility of Google’s algorithm updates.

Thom Ives (Ph.D.), Lead Data Scientist at UL Prospector, likened data to crude oil that needs to be refined and cleaned. He warns that data “could be dangerous when handled the wrong way.”

This makes decision-makers nervous.

As it turns out, even though 74% of companies agree they want to be data-driven, according to a report by Forrester, only 29% could act based on analytics results.

As much as data-led decisions are excellent in growing business, the mistakes can be devastating. Perhaps the possibility of making mistakes deterred the bulk of the other 71% who make decisions by experience or gut-feelings, or simply follow the status quo.

Often, this happens at the expense of tapping into our 59 zettabytes of data (that’s 59 followed by 21 zeros!) to generate business-transforming insights.

Peep Laja, CEO of Wynter, aptly summarizes it, “We’re data-rich, but insight poor.”

10 Actionable Steps to Turn Data into Insights

Metrics are easy; insight is hard — it’s one thing to gather lots of data, but another to make them a valuable asset. Thankfully, there’s a tried and tested method to get answers.

Cue in the scientific method.

This is not a eureka moment, though. Scientists have been using this method when deriving insights from data for centuries.

The 10 steps we’ll show you draw inspiration from the scientific method and pave the way to actionable insights and recommendations.

Turn data into insights steps
The age-old scientific method is key in obtaining insights from big data

Here’s a 2-minute summary of the 10 steps to turn data into insights.

Let’s jump right in:

1. Start With the Right Questions

Asking the right questions before you dig through data ensures you don’t spend time on the wrong things.

It’s like setting a clear destination before embarking on a trip.

Before you comb through mounds of data, figure out what questions you want your data to answer. That way you avoid coming up with insights that have no impact on business goals.

For a SaaS company, some questions to start with are:

  • How many blog post readers moved on to other pieces of content?
  • What percentage of my website traffic fits my buyer persona?
  • What stage of the sales funnel leaks the most?

2. Track the Right Metrics (No Vanity Metrics!)

Insights that steer the business in the right direction do not come from staring at the wrong metrics.

Vanity metrics, especially. They make you look good but do not add to your insight framework. Example: page views and number of clicks.

Besides, the wrong metrics can be distracting. Since you’ve decided on a question that needs answering in step 1, pinpoint the metrics you should be tracking.

Aniekan Inyang, a Data Scientist at Stears Business cautions against not accounting for industry-specific nuances:

This can lead to choosing a wrong feature to track a metric, not tracking a relevant metric or interpreting it wrongly.

Use that to beat a path to a hypothesis you can test.

Speaking of hypotheses, have you tried our A/B testing hypothesis generator yet? Use our free hypothesis generator tool or learn more about building a hypothesis.

3. State Your End Goals

You most likely have certain business goals pre-test. These have to be closely aligned with your test objectives.

From the questions you started with, you figured out what you want to track. But what are you aiming to achieve with this?

Write this down as it helps you develop a specific, measurable hypothesis.

4. Integrate Your Data Sources

The data sets you have are only a section of the population and won’t always tell the full story.

Dr. Thom Ives shared:

It may have biases we don’t know about and is going to be weaker than all the data.

The more actionable data you gather, the closer you get to accurate stories.

Your data interpretation hits closer to the bullseye when you bring all your sources together. Make sure you use the right tools to integrate disparate sources so you don’t miss out on gathering meaningful customer insights.


Run your tests with an A/B testing tool that plays well with other software. Convert Experiences integrates with dozens of tools that may be in your tech stack.

5. Use Context and Visuals to Simplify Data Sets

Visuals are pretty common with data today. You hardly ever encounter an incomprehensible raw form of data. Yet, without the right context, you’re either not getting the full story or getting the wrong one.

For context, dissect your data using the 5 W’s:

  • Who (audience, leads, prospects)
  • What (goals, events, observations)
  • When (timeframe, schedules)
  • Where (webpage, social media, landing page), and
  • Why, (why did it happen?)

Context makes your data jump off the screen with more meaning behind it. It reduces the chances of making a mistake.

Added to accurate visuals, those chances get even lower. But errors are made on visuals too.

For example, it’s common to make a costly mistake with bubble charts. Varying the radius instead of the area of the bubble to corresponding values leads to inaccurate data storytelling like in the picture below.

Visualization turn data into insights
Source: The Next Web

Let’s use the orange bubble in the top-left and the green one next to it to emphasize. The orange bubble looks 4 times larger than its green neighbor.

Without the actual values labeled inside, this can be misleading. The orange bubble’s value ($1.84B) is only 2 times that of the green ($0.92B).

Here’s a funny blunder by Fox News:

Visualization blunder example
A pie chart that adds up to 193% (Source)

6. Segment Your Data

Cutting up data into segments can help you make better sense of it. Google Analytics, for example, has built-in features that make this easy to do.

Divide web traffic according to certain similarities, and it’ll simplify the process of extracting insights. Segmentations can deepen your understanding of your target audience.

Also, when segmenting, think beyond the old school age and gender segments. There are much more details with which you can group web visitors.

One way to do this is segmenting customers by transactional worth (value segmentation) — that is, how much they’re likely to spend on products. You’ll have to use past transaction data to achieve this. Data such as how much they’ve spent, how frequently they spent it, and the value of the products they bought.

Once you experience this simplicity once, it quickly becomes a staple process in your insights strategy.

Here’s another example that illustrates the importance of proper data segmentation:

7. View Data in the Right Time Frames

Making decisions based on insights drawn from a slice of time can be disastrous. Only looking at the little picture with zero references to historical data is a common error.

Data usually has a backstory.

It’s important to check that out to make sense of the present. Sometimes events happened in the past in response to external influences like holidays, seasons, economic cycles, etc.

Take this into consideration as you explore the full spectrum of a trend to get a more accurate read on things.

8. Spot the Right Patterns

Climbing and falling — two of the easiest trends to observe on a line graph. This is usually how page views and engagement data are displayed on GA.

Other types of plots such as time series and scatter plots help us picture patterns in data. You can spot when there’s an upward or downward trend, visualize a correlation between two variables, and more.

They’re all tailored towards revealing the stories behind the data. A word of caution: never view patterns in isolation of their context.

data patterns in context

In analyzing your plots, MIT Professor Dr. Rama Ramakrishnan suggests matching your plot with preliminary expectations:

Is there anything that doesn’t match? Anything that makes you go ‘That’s odd’ or ‘That doesn’t make any sense.’? Zoom in and try to understand what in your business is making that weird thing show up in the data like that. This is the critical step. (…) You may have just found an insight into the business and increased your understanding. Or you may discover that there’s a bug in the way your data has been collected or calculated (Twyman’s Law).

9. Craft a Winning Hypothesis

When you’ve analyzed your data and drawn accurate inferences, it’s time to come up with a hypothesis you can test.

In crafting a hypothesis, you are figuring out a solution to a problem that you can verify with experimentation.

A measurable hypothesis consists of 3 parts:

  • Observation,
  • Execution, and
  • Outcome

Here’s a real example coming from a Convert partner:

Observation: From analytics data, we observed a high bounce rate on our flagship product’s page. We also carried out surveys, polls, and usability research and found out that users didn’t understand the value of our product and trust it. Also, most visitors didn’t scroll further down the page.

Execution: We want to add better copy to the fold area to retain more page visitors, address the trust issues, and boost conversions on the page.

Outcome: This should lead to more web visitors scrolling through the page, desiring our flagship product, and buying it. We will measure this by lower bounce rate, higher conversion rates, and revenue.

Once you’ve landed here, the next step is testing.

This example is an actual hypothesis that led to impressive results. For more details about the experiments, check out the first actionable insight example below.

10. Prepare to Experiment

With the hypothesis above, you can do what expert conversion rate optimizers do and run a test.

Up until this point, your hypotheses — even though they’re borne from data — are only as good as intuition.

Experimenting gets you closer to creating a rock-solid fact.

This is where you start getting the ROI for your data analysis.

The scientific approach helped us turn our raw, incomprehensible data into something readable. Then we applied the power of data analysis to unveil the juicy insights it contained.

We developed measurable hypotheses from these insights and took the next logical step: experimentation.

There are hundreds of tools that take us through these steps. But Convert ties them all together at the end and brings us to our ultimate goal — actionable insights.

Convert integrates with dozens of tools so you can experiment and discover incredibly powerful insights that directly lead to higher revenue.

Review Convert’s Integrations

Going Beyond Insights: How to Turn Insights into a Strategy

Insights aren’t useful in achieving business goals if they aren’t translated into strategy and acted on.

How can you actually use the insight you get to drive positive benefits that directly influence your organization’s bottom line?

Let’s share 3 examples:

Turning Data into Actionable Insights Examples

BestSelf Co. discovered a leak on their flagship product’s page. So they worked with SplitBase to plug it.

How did they do it?

Using various means to gather data such as polls, surveys, and heatmaps, they found the culprit.

The benefit of the product wasn’t communicated well enough, so people weren’t even getting past the fold area. From this, they crafted the hypothesis we shared earlier.

They ran a test and found they were right. The new headline clearly stating the major benefit of the product and social proof significantly boosted the sales of the product.

Case study #2: Was this free shipping promo making or losing money?

This was the question on the minds of the team running a luxury handmade glass eCommerce store.

They launched a free shipping promotion and found an increase in conversion rates. Although that meant more money, considering the costs of shipping these products to customers, was the offer enough to offset shipping costs?

Now, how they found their answer…

They called on Brave One, a conversion rate optimization agency,  who came up with a plan to find out if they were losing or profiting, and by how much.

With Google Analytics and Mixpanel for gathering data and Convert for experimenting, Brave One compared the site without the offer to a version of it with the offer.

Running the business with the offer brought in $16,000 more than running it without in the same time frame.

Case study #3: Data helped Nike adjust their targeting and touch the hearts of a wider audience

When Nike wanted to launch a campaign called ‘Find Your Greatness’ at the start of the 2012 Olympics, they dug into their data and found this:

Most of their target audience wasn’t professional athletes. They were people who admired the pros and wanted to be like them.

What did they do with this?

They adjusted their targeting.

Nike usually goes after the pro athletes. But this time around, they decided to inspire everyone regardless of their fitness level to push their limits.

One of the videos of the campaign had more than 3 million views.

And it doesn’t stop there: Adidas spent millions of dollars to get an Olympic sponsorship, yet Nike enjoyed that same level of exposure with less than half of that marketing budget.

This video from Convert’s first ever video course explains the three types of data insights and how to determine if they are valid.

Summing Up

Experimenting shouldn’t mark the end of your optimization journey.

It should be a continuous process because we are never always spot on with our insight.

Also, remember… in case you lack the skills to conduct in-house data analysis, you can always rely on the know-how of data experts.

Dr. Thom Ives suggests that as more data comes in, we have to refine the inferences we made with the old data.

And the good news? This way, we continue to approach insight that’s more representative of our audience and make much more precise predictions and decisions.

CRO Master
CRO Master
Originally published February 24, 2021 - Updated June 06, 2023

Mobile reading?

Scan this QR code and take this blog with you, wherever you go.

Uwemedimo Usa
Uwemedimo Usa

Conversion copywriter helping B2B SaaS companies grow.

Carmen Apostu
Carmen Apostu

In her role as Head of Content at Convert, Carmen is dedicated to delivering top-notch content that people can’t help but read through. Connect with Carmen on LinkedIn for any inquiries or requests.

Start Your 15-Day Free Trial Right Now.
No Credit Card Required

You can always change your preferences later.
You're Almost Done.
I manage a marketing team
I manage a tech team
I research and/or hypothesize experiments
I code & QA experiments
Convert is committed to protecting your privacy.

Important. Please Read.

  • Check your inbox for the password to Convert’s trial account.
  • Log in using the link provided in that email.

This sign up flow is built for maximum security. You’re worth it!