This article was originally published on Online Metrics (June 7th 2016).
"When you have mastered numbers, you will in fact no longer be reading numbers, any more than you read words when reading books. You will be reading meanings."
~ W.E.B. Du Bois
Numbers and meanings are similar to data and actionable insights.
Data is great, but data by itself is not enough.
Big data is awesome, but it can easily lead to data overload.
We live in a world where the number of data sources available is growing every day.
You need to find the best way to make your data insanely useful.
I have been busy connecting to over 100 Analytics Experts around the globe, making a mess of my email box and in spending dozens of hours to collect the answers and compile the post.
An AWESOME experience! :-)
It doesn't matter whether you run your own business, are a digital marketing consultant or analyst. Either way you will benefit from reading and implementing the awesome tips and strategies from the experts shown below:
Alex Clemmons | Amir Tohid | Andre Mafei | Andrey Osadchuk
Andy Crestodina | Ani Lopez | Annemarie Klaassen | Ben Adams
Brian Clifton | Carlos Escalera | Charles Farina | Chris Meares
Damion Brown | Daniel Waisberg | David Kamm | Dean Levitt | Dominic Hurst
Doug Hall | Egan van Doorn | Eric Fettman | Eric Siegel
Franck Scandolera | Himanshu Sharma | Jacob Knettel | Jean-François Bélisle
Jeff Sauer | Jim Gianoglio | Jim Sterne | Joel Davis
Jordan Louis | Julian Jünemann | Julien Coquet | Kevin Anderson
Lea Pica | Manoj Jasra | Marco Pasin | Mike Sullivan
Mikko Piippo | Quentin Laveau | Şahin Seçil | Sameer Khan
Sayf Sharif | Simo Ahava | Stéphane Hamel | Tim Wilson
Todd Belcher | Tom van den Berg | Yehoshua Coren | Zorin Radovančević
They have all answered the following question:
“What is your number one strategy to turn data into actionable insights?”
Without further ado, here are the experts…
Alex Clemmons from Cardinal Path
"It all starts up front: on day one of the project everyone should ask themselves "what questions do we need to answer in order to succeed?" From there you can determine the data you need to collect to answer these questions. Additionally, since you're asking this question on the first day you give yourself enough time to properly capture that data before your project launches.
Build a measurement strategy to capture this and then treat it as your analytics roadmap through the entire project. That way once it's time to actually run your analysis you'll have a blueprint for where to start and have the confidence to know that you already have good, clean data to base your hypothesis on."
Amir Tohid from Analytics Effect
"In my opinion, first of all companies should have a clear understanding of what they want to achieve and what are their business goals. My approach is to always start with “small data” which makes it easier to produce insights fast. Filtering, grouping and segmenting the data plays a very important role.
During analysis stage, I always focus on trends, not data itself. The best insight comes from looking at trends especially when they change direction and I compare time ranges such as week over week, month over month etc. Moreover, I also search for strong relationships between variables or correlations.
This is a practice and this has to be done over and over again to improve the desired outcomes."
André Mafei from Upmize
"My strategy to turn data into actionable insights is to integrate data sources for better and faster business decisions.
Data warehouses should do this, but one of the problems is that to integrate new data sources and to build new reports it is necessary the help of IT, what causes overload, then teams quit waiting and create their own solutions, ending in several "data silos" (nice name for "data mess").
Another problem is the cost for large amounts of data, what made online companies like Google and Facebook invest in creating new big data solutions like Hadoop.
So the solution is to integrate traditional databases and big data solutions into a unified place where you can control:
1) Collection (integrations, elasticity),
2) Management (data quality, governance),
3) Analysis (data exploration),
4) Data solutions (automation).
Tip: research about data lake and CDO (Chief Data Officer)."
Andrey Osadchuk from BizTech Enterprise Solutions
"Make sure that the insights are aligned with the business requirements and primary KPIs of the decision makers. The analysis should be focused to help businesses reach their goals. Work on a single problem at a time. Build a story and talk to those whose KPIs it affects. While for one audience a positive expression like “you will get extra $ if” works best, for the others consider a negative scenario “you will lose $ if”. Start with small actions, break complex ones into pieces and proceed with them sequentially. The insight can be called actionable not earlier than the first action is completed."
Andy Crestodina from Orbit Media
"First, don't confuse reporting for analysis. Looking at reports isn't the same as finding insights and taking action.
This is what reporting looks like...
The line goes up, you smile! The line goes down, you frown.
But analysis means more. It means...
- Asking questions and finding answers in the data.
- Forming hypothesis, testing and measuring results in the data.
- Understanding the meaning behind the numbers and lines and taking action based on that new understanding.
Most people who say they "use Analytics" are really just looking at reports."
Ani Lopez from Dynamical.biz Consulting Inc
"Obviously 'actionable' is the key here, especially when 'insights' is such a buzzword in the digital analytics industry. Data can be transformed into many things, shiny useless things for the analyst's self indulgence or solid profits for the company.
To make data actionable, business goals have to be the true north in your compass while working at any given point of a measurement strategy but more importantly when you dive under an ocean of data trying to come up with hypothesis to validate as it's easy to overlook the context and the objectives.
In practical terms, print them in big bold letters and place that paper somewhere in front of you.
Different questions are if stakeholders provide clear business objectives and if they have the vision to take action upon analysts' recommendations but that's a nightmare to debate another day."
Annemarie Klaassen from Online Dialogue
"My number one strategy to turn data into actionable insights is to ask yourself what you are really looking for. If you don’t have a clearly defined and specific question you certainly won’t get a clear answer either. If you ask generic questions, you will get generic answers which aren’t very actionable. Ask specific questions and you will get specific actionable data!
Say for example that the question you need to answer is: “how well is our website performing”?. This is a very generic question and being an analyst you’re prone to dig deep into the data and come up with all kinds of metrics and graphs: you make a graph of the bounce rates over time, you analyze the page load time, the number of conversions (micro and macro), the revenue per product category per traffic source and so on. All the metrics indicate some form of performance, but none will tell you the exact answer. A better actionable question would be: “which product categories are underperforming in comparison to last year and can this be explained by certain traffic sources or marketing campaigns?”
Ben Adams from The Wharton School
"What is the question you're trying to answer? As a data scientist, you should always start out with a question. Once you have a question (or set of questions) then ask yourself this: how is my behaviour going to change - how is my organization's behaviour going to change - as a result of the answer? It may be great to say “75% of our customers are on the East Coast” but are you going to do anything differently knowing that number? How about: “75% of our paying customers are on the East Coast, and we send 50% of our mail to the West Coast.” Now you’ve got a behaviour change ready, based on data."
Brian Clifton from Successful Analytics
"A company’s ability to satisfy the needs of a website visitor depends on two important factors:
1. Visitor expectations, discerned from how they got to your content— what search engine, campaign ads, or social conversation drove their decision to seek you out.
2. User experience, how easy it was to use your content, to navigate around and to engage with you (contact you, purchase, subscribe, give feedback).
It is your organisation’s ability to manage, analyse, and improve these two factors that determines your digital success (or not). The key is to think in terms of insights—not data.
When Paul came to me with his question, I thought long and hard about how to answer it. Essentially there is no single strategy or silver bullet that an organisation should focus on in order to gain insights (its why I write books on the subject!)
Producing insights requires an understanding of your business and its products, your value proposition, your website content, its engagement points and processes, and of course its marketing plan. Your analytics tool provides the data (and lots of it) that enables you to assess these. However, people—not machines—build insights. Smart people are required to sift through the noise to find the useful data, translate it into information to explain what is happening, then build stories of useful knowledge for the organisation—the insights.
This process is a detailed one by necessity. Building an environment where you can trust your data, understand it, and make important decisions based on it requires a deep level of immersion, not a superficial scan of reports. It also has a breath that, if it is to be successful, it must involve numerous people at senior levels of your organisation. In other words, it's a "team" effort. In most scenarios this means combining your internal business experts with the expertise of external analytics specialists."
Brian Clifton (PhD) is the best selling author of Successful Analytics: Gain Business Insights By Using Google Analytics and the series Advanced Web Metrics with Google Analytics. He was Google's first Head of Web Analytics for Europe (2005-8) and built the first pan-European team of product specialists. A legacy of his work is the online learning centre for the Google Analytics Individual Qualification (GAIQ).
Carlos Escalera from Ohow.co
"The key for turning data into actions is learning to interpret what it is trying to tell you. For instance, one of the (many) steps of my daily strategy is the deep analysis of the bounce rate. This critical and often misinterpreted metric, besides helping me measuring experiments and improving the user experience which are common scenarios for any analyst, also had helped me find broken code on my site or even generate new ideas for posts. A couple of examples:
If you're a fan of plugins, especially if you like to tweak them as I do, occasionally things can break, due to updates or incompatibility with new plugins. If the broken plugin is not used widely on your site, it is hard to detect it. Here is where the bounce rate comes to the rescue; sudden changes can help you detect broken parts on your site and repair the damage to stop the leak of readers/customers.
Now, bounce rate and generating new ideas for posts may sound weird, but it actually works. When an old post starts leading more visits than usual out of nowhere but at the same time the bounce rate increases, it means that initial purpose of the post doesn't fulfill the need of the new readers. This behavior instead of a problem frequently indicates an opportunity, sometimes to update old content or create a new post where you can redirect that audience.
Just recently, I used these in one of my articles talking about the last year update of mobile friendliness, on which I only had to adapt the introduction to get people looking for fresh information about the topic to stay and read.
In these cases interpreting the data, rather than just seeing the numbers as more, less, high or low, helped find opportunities that otherwise would have been hard to see."
Charles Farina from Analytics Pros
"In order to drive the action in “actionable insights” it’s critical that your data and insights are relatable. Many analysts are unable to drive action, because they never tie their findings to business impact. Insights start with the question you are answering. That question is the key to driving action. Start with this question, measure the loss/gain caused by your finding, and then answer the question by recommending your actions. The easiest path for most analysts to actionable insights is through a solid optimization program they can plug into."
Chris Meares from MaassMedia
"Data by itself does not lead to actionable insights. First one has to understand the business questions that are trying to be solved as on organization. This could be as simple as which marketing channel drives the most revenue or as complex as measuring the lifetime value of consumers that download a company's app. Once the business questions are known, we can then understand which specific data needs to be collected and analyzed. Through this data analysis we can begin to create hypotheses to answer specific business questions and test our hypotheses to gain actionable insights."
Damion Brown from Data Runs Deep
"I’m going to kind of steal an idea from Simon Sinek and say that for truly successful insight, you need to Start With Why.
There’s a temptation as an analyst to focus on things like new report features or stuff that’s “interesting” – and as many have said, the word “interesting” is probably the most pointless word in the web analyst’s lexicon.
By starting with the Why, you’re aligning the process to the organisation’s most important outcomes, and you’re not getting distracted investigating tangents and disappearing down rabbit holes."
Daniel Waisberg from Online Behavior
"Make it interesting. As with anything in life, boring stuff is more likely to be ignored. A smart analysis can have powerful and actionable insights, but if it is not presented in an interesting way it won't be "heard". I've written about creating data stories in the past, and with the launch of Data Studio for everyone, that became easier and more effective! So next time you find yourself creating a report full of tables and pie charts, think a little harder and try building clear visualizations that communicate the data in a more understandable way."
David Kamm from iBeam Marketing Consulting Services
"My top strategy and guidance for turning analytics data into actionable insights is to be very clear about how specific metrics being tracked contribute to one or more business goals of the organization. If these connections to desired outcomes aren't clear to all involved, and/or aren't direct enough, then the movements in the data won't mean much and these metrics will just contribute to 'analytics overload'.
When metrics are both insightful and actionable, the team monitoring them can learn something new by watching the data; something they didn't really know before. And the team can make real-world marketing adjustments based on the data (e.g., new content or campaign changes), with reasonable confidence that these changes will impact the metrics that matter.
So it's about measuring the right things based on business goals, and understanding how potential changes are likely to drive these metrics in a positive direction."
Dean Levitt from Teacup Analytics
"All analysis can be boiled down into three possible possible actions. You could do nothing, for example if a channel is insignificantly small. You could optimize, if a channel is not converting well. Third, you could grow a channel that is performing well.
Take a look at any channel or segment and you'll notice that everything falls into one of these three categories. Assuming a channel doesn't fit in the "do nothing" group, once you know whether to optimize a channel or to grow it, deciding on the right action can be as simple as using Google to ask, "how do I grow organic search."
Dominic Hurst from dominichurst.com
"For me the key in turning data into actionable insights revolves around setting an underlying digital goals framework from your organisations objectives.
Straight away you can focus your time and effort on dimensions and metrics that really matter. No more top pages or aggregated metric dashboards that are meaningless never mind insightful.
The extra time comes in handy though as you now have the time to dig deep, segment and filter your data, uncovering the true insight behind a number.
But it doesn’t end there. Because you have a goal at stake, a goal management is aligned too; your insights have impact and become actionable. For example you might uncover that a drop in a goal is linked solely to mobile and fell on a certain date. Working with developers you can find out how a recent release broke the responsive view. Now showing the metric and this insight gives management an immediate action to take."
Doug Hall from Conversion Works
"Remember, you're working for the user.
It's so easy to get into a cynical and wrong frame of mind:
- What's converting?
- How do I get the user to do what I want?
- How can I better monetise users?
Yuk. No. Wrong! Users aren't meat that you shove into the top of a grinding machine to produce conversions. Users are your customers and your site needs to deliver what the user wants.
Great sites aren't great because they make loads of money, great sites make loads of money because they are great.
Users are the judge of the greatness of your site so your data is your users telling you how to make your site better for them. Embrace these insights, deliver for your users first and business success will follow."
Egan van Doorn from eganvandoorn.nl
"Choose your metrics wisely, make sure that they connect with the goals of your organization and educate everyone working with the data how these metrics contribute (or add up) to the main KPI. The ones receiving the data should be able to influence the metric within their responsibilities for the site, marketing or product. Next up, always... ALWAYS report against target. Setting up targets for every metric in your KPI chain indicates that you thought about the implication and attribution of a specific metric in reaching your overall organization goal(s)."
Eric Fettman from E-Nor
"People: invest in people who have passion for both analytics and end-user experience. While our industry has evolved tremendously, and while the new generation of marketers is now more data-aware than ever, we're still not dedicating enough resources to turn data into action. Organizations that are achieving positive impacts are those that put data-driven, customer-focused problem solvers behind all digital transformation initiatives.
Process: closely aligned with the previous thought, and as is true in so many industries, process and consistency are what keeps the lights on. At the implementation phase of a Web or mobile analytics platform, are you completing the audit and QA necessary to address all potential gaps and flaws in your data capture and integration? For inbound email, SEM, banner, remarketing, paid/non-paid social, and SMS campaigns, are you taking the little bit of extra time necessary to tag your inbound links, and in way that populates your acquisition data as a smooth, unfragmented, understandable hierarchy? At even a more basic level, are you maintaining a timeline with specific dates for these marketing campaigns, as well as for design and development changes and fixes, planned (or unplanned) downtime, and any other factors that you need to remember so you can coherently interpret your data one or two or six months after the fact?
Success: it's a worthy mantra: focus on success. Define your success early on, re-evaluate your definition of success periodically, and make sure to focus your analysis on success. At which stages are users dropping out of your conversion funnel? Which marketing channels are generating the greatest Ecommerce revenue, both for last click and assists? Consistently drive towards improvement for a small and very deliberate set of KPIs, through analytics, testing, and also qualitative inputs: this is the catalyst for long-term, bottom-line benefit."
Eric Siegel from Predictive Analytics World
"The most actionable win from big data is predictive analytics, since each of the millions of per-individual predictions it generates directly inform the treatment or action taken towards that individual - such as whether to call, mail, approve, test, diagnose, warn, investigate, incarcerate, set up on a date, or medicate. This is the form of information technology that's transforming all the main activities organizations do, bolstering the effectiveness of our largest-scale operations.
"Predictive analytics applies across sectors and functions - it targets marketing, streamlines manufacturing, drives fraud prevention, improves financial decisions, optimizes social networks, empowers spam filters, fine-tunes law enforcement investigations, improves healthcare decisions, and optimizes political campaign activities."
Eric Siegel is the author of Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die (www.thepredictionbook.com).
Franck Scandolera from webAnalyste
"As an expert on data implementation, my first strategy to turn data into actionable insights, is to collect actionable data, like the most useful attributes of the traffic source, of the content, of the product, of the customer, of the visitor, of the functionality to analyze the factors (independent variables in predictive analysis) and the segments (characteristics of cluster analysis) that affect the conversion.
Foremost, the actionable data should be valid and reliable, the measurement of the phenomena must be correct, and that one time or thousand times in the context given.
As an expert on data analysis, my first strategy to turn data into actionable information is to get a good understand of the business answers researched. Once you are at the end of the story, it's easier to determine the best data set (dimensions and metrics), the appropriate techniques for analysis (simple, transparent and replicable) and the best visualization to tell the story of "what is important and why you should care".
In summary, actionable insight = good understanding of the business questions + good, valid and reliable data set + good data visualization + good story."
Himanshu Sharma from Optimize Smart
"I have no number 1 strategy as such. I follow a process. In that process I create strategies, which vary from business to business. So there is no one size fit all strategy. Moreover, I do not use strategies to turn data into insight.
For that I rely on my data interpretation skills (maths and statistics). My strategies are largely geared towards ‘getting things done’ (like convincing the client for carrying out a test) which I think is the most important and challenging aspect of providing consultation and solving customers’ problems.
Everyone has got data and everyone has got their own personal data driven insight (aka opinion). Different people can interpret the same data differently. It all depends upon the context in which they analyse and interpret the data. The one who has got superior understanding of context, will interpret the data more accurately."
Jacob Knettel from PFSweb
"The key to turning data into actionable insight comes from knowing the true business goals for the client or brand you are working for. When the analyst knows these goals (whether macro or micro), he can focus you his measurements, insights and recommendations around improving these goals. This creates a beautiful harmony between the analyst and end-user, where the analyst feels empowered because he/she knows that what they are researching is going to be meaningful for the business, and at the same time, the end-user gets the information they actually need to make smart business decisions.
Without knowing the business goals, an analyst can spend hours mining for gold, when at the end of the day, what the business really needs is silver."
Jean-François Belisle from Flatbook
"The number one strategy is to truly understand the “customers behind the data”. Thus, you need to be fully-aligned with the business unit experts. You need to meet them, to lunch with them and to challenge them. They are the ones who have the best “feeling” about the customers that generate your data since they are the ones who interact with them. This is why client representatives are often a golden mine of information about a company’s customers. They are not data experts, but they are the ones who can tell you why you have some strange results and then you can take these insights into account for your analyses. To conclude, “truly understanding the customers behind the data” is what makes the difference between a good analyst and a great analyst and in many cases it makes the difference between “being listened” versus “being indispensable” in an organization."
Jeff Sauer from Jeffalytics
"The first thing I do with any set of data is establish context with what I am seeing. What do these numbers mean? Are they important? Does it affect the way we do business? Once this is established, I use the important data to establish our baseline performance. This is the stake in the ground for how we have been doing to this point.
Now, as a marketer, my focus turns to growth. Growing beyond our baseline to continuously improve upon these efforts. This involves setting targets and establishing a plan for getting more. Then it’s all about working the plan (and checking back with your data to make sure things are working as planned.)
While I don’t really use the term actionable insights anywhere along the way, this entire process of continuous improvement can be summarized as using data to take action."
Jim Gianoglio from LunaMetrics
"The biggest problem I see in organizations is the assumption that having a lot of data will automatically reveal a panacea. They think if they can figure out how to manipulate the data in just the right ways, or use the latest big data tool or service, then actionable insights will just begin to appear.
The best strategy to get insights comes before the data collection, tool selection, or analysis. The number one strategy to get insights is to ask the right questions! Exploratory analysis is important (and fun!), but trying to get insights that way is like going to the grocery store hungry and without a list. You end up putting everything in your cart, whether you need it or not. If you have a focused, business-critical question that needs to be answered, then that informs the data collection, analysis and visualization that will reveal the most important insights."
Jim Sterne from Target Marketing
"If I can only pick one strategy to, it would be to clearly and deeply understand the problem to be solved. A talented tea leaf or tarot card reader will engage you in conversation and use the tea or the cards as a tool to find out what you feel is important - and have you create the 'fortune'. The difference with analytics is that there actually is something of value inside the crystal ball. That information has value when it is turned into insights based on the conversation with the 'client'. Those insights will be actionable because the client will share ownership."
Joel Davis from Google Analytics Demystified
"Don’t focus on or obsess about any one particular metric. While each individual metric has the potential to provide important insights and direction for strategic decisions, it is only when you understand the overall pattern of responses that you obtain the deepest insights. One way to uncover overall patterns and interrelationships among metrics is to view each individual metric as if it were a piece of a jig saw puzzle... How does the insight from each metric fit with what is learned/revealed from other metrics? How does simultaneously viewing two or more metrics provide a different perspective than viewing each metric individually? Is the pattern of response the same or different across different but related metrics? Finally, once the pattern across metrics is established, the last – and most important - task is to identify what pieces are missing and determine why the observed pattern is occurring."
Jordan Louis from jordanlouis.ca
"When you’re looking to turn data into actionable insights, you’ve got to make sure you’re aware of the context within which that data was collected. You can’t take action on data if you don’t know what it means or why it is significant. If I’m reporting and analyzing data from a national census, I’m going to treat it differently from sales and advertising data from an e-commerce website because the data in each case is talking about different things. However, you can and should use one dataset to inform the action taken on another, such as using the average income by postal code from the census to better target your advertising efforts, then measure the effect this has on sales per advertising dollar. Just be sure not to get confused: data will tell you different things depending on where it’s coming from. A smart analyst appreciates the context."
Julian Jünemann from MeasureSchool
"My number one strategy: Applied segments. Digital Analytics comes down to segmenting, segmenting and segmenting again to understand the user behaviour. Once you have identified a profitable segment it is time to take action. Retargeting and targeted Emails are my first choice to get from data to action. It is simpler than ever to user your GA advanced segments to build a Remarketing list or tag your email subscribers with the help of Google Tag Manager. The positive effects will be visible to the bottom line in no time and you will have proven the ROI of your analysis."
Julien Coquet from juliencoquet.com
"The first strategy I apply is the actual data collection strategy, which is too often overlooked. This implies playing the role of the "analytics midwife" in workshops with digital marketers in which the goal is to get them to express their measurement requirements: what kind of information they want to collect about their content, campaigns and visitors.
This may sound like Analytics 101 but in reality most marketers usually start collecting useless data before resorting to contacting me because their data is rotten.
But that doesn't stop there: once technical partners implement your tagging, you need to verify that the decisions you make based on digital analytics data are supported by quality data - otherwise you will fail even harder.
The best way to ensure your data is collected properly is to apply data quality processes for digital analytics quality and automate data verification as much as possible."
Kevin Anderson from ING
"A succesful analytics program needs a solid KPI framework. A KPI framework gives an overview of what the organization is trying to achieve and ties that to how we will be measuring this. This exercise will push all investments in people, tools and processes in the right direction. And most importantly it will need a continual dialogue between management and data analysts. So my number 1 strategy is: start with a set of KPI's."
Lea Pica from Leapica.com
"The #1 tip I have for turning simple data into actionable insights is to go the extra mile to answer all of your stakeholder's questions along the insight spectrum. This means more than just plopping numbers in a chart, slapping a title like "Campaign Results" on it, and walking away. Nope.
This means developing a keen intuition of your stakeholder's aspirations and challenges. Visualizing the data in a way that promotes cognition, not confusion. Clearly articulating your specific data story with as much what, how and why behind it. And finally, extending your added value by attaching recommendations that are aligned with the data story, assigned to a specific owner, and time-bound for accountability.
With this framework, you will supercharge your credibility and indispensability that go well beyond crunching numbers at your desk!"
Manoj Jasra from Shaw Communications
"I will answer this question with a slightly different lens: Having worked in large organizations over the past 9 years I would have to say the key is to build a culture that sees data as the epicenter for their digital strategy. Without this it is very difficult to consistently prioritize the value of data analysis in order to achieve actionable insights. In order to build this data-driven culture it takes: hiring the right talent who can decipher data and are passionate advocates, having a sound analytics implementation and finally championing/educating the value of insights and their impact to all levels within the organization."
Marco Pasin from Analytics for Fun
"Segmentation! By grouping together visitors/clients that have some attributes in common I can really start digging deeper. Choosing which segments to study of course it depends on what is the business question I am trying to answer (I always define it clearly before starting any analysis!). Tools like Google Analytics have some powerful segments already built in (mobile vs desktop, converters vs non, etc.), or you can set up your own. And most data is ready to be analysed. But sometimes things get a bit harder like when I have to join online behavior data with customer databases or when trying to solve more complex data science problems. In these cases, before performing any segmentation, I need to design, clean and prepare the final dataset. Cleaning data is a critical step without which you might ruin your entire analysis.
Using meaningful data visualizations allows me to spot patterns and get insights on segments faster than just looking at tabular data. I use a lot of viz to explore, understand and present segmented data.
If you want to take action on your data, go for segmentation!"
Mike Sullivan from Analytics Edge
"Start with a hypothesis and a willingness to change if the data validates your hypothesis. Too many people jump into an analysis and expect the data to 'talk to them' if they look at it long enough, but if they actually see something, they aren't willing or able to do anything with the observation. Start the analysis when you are looking for something to change or improve, then segment by as many different dimensions as possible. Note all your observations, even the negative or neutral ones. Most importantly, act on the results."
Mikko Piippo from KliKKi
"A strategy I use very often is simple: First I split the data using a primary and a secondary dimension. In the second step I visualize the data as a heatmap. Very often this is enough for getting something out of a large matrix of numbers.
This strategy - or method - can be used in countless ways. For example, create a colored heatmap with days of week on the X axis and hour on the Y axis. Fill the matrix with conversion rates or number of conversions.
Excel and Google Sheets are the easiest tools for creating heatmaps. The visualizations are not pretty, but they are more than adequate for exploratory analysis and convincing the client."
Quentin Laveau from Booking.com
"Always start by segmenting your data, it will give you the “WHERE on your website users are leaking” and “WHO are those users” (i.e: Returning users coming from Paid Search landing on Product page). Raw data will become Information once you understand the context and relationships between different segments.
Then, back up your clickstream data with qualitative analysis (session replay, heat maps, customer feedback). It will help you understanding the “WHY they are leaking” and “WHAT you need to change”. You have reached the Knowledge step, when your information has meaning and purpose.
Now, synthesise the knowledge by writing a hypothesis: If ___Then ___ Because ___. That’s the Understanding step, where you’re now able to take actions based on your data."
Şahin Seçil from Frosmo Ltd.
"As you know, "Data is only valuable if you can translate it into actionable insights." On this point, my question is: Which report can give actionable answers about conversion?
The answer is a "landing page report." It sounds easy, right? You don't need to create complex reports to start to define your customers. Cohort reports, CLV reports, User-ID reports, etc. can give valuable information but you should be an expert in creating and analyzing these reports.
Let's start with checking out the top 10 top landing pages. If some of the landing pages' conversion rates are lower than average, you should think that the users can’t get enough of the information that they are looking for. Your content can be related to your products or services, but maybe your digital marketing campaign targeting is not true. For instance, if the Adwords campaign targeting is not related, your landing page conversion rate will be worse too, or your website modules aren't as useful as you think. Or maybe you can consider showing different content for returning visitors. Consequently, you can segment these users by checking their past behaviour.
So you can start digging into your data by looking at the landing page report to improve your KPIs."
Sameer Khan from KeyWebMetrics
"Dataset, numbers, tables and charts are all passive signals. The story behind the data is where it all connects together. The story is what drives, motivates and inspires people. I like to invest time in developing meaningful stories from data so we can drive actions across our team and cross-functional groups in the organization. Also, always remember to provide context for each dataset before sharing reports and presentation. It will save ton of emails back and forth and prevent data-confusion."
Sayf Sharif from Seer Interactive
"It’s imperative to understand your audience. All the other aspects of your data from acquisition, to behavior, to conversions; all revolve around the users, and how you can segment based on your specific audiences. That means digging past the standard dimensions like city or source, and defining your own custom algorithms to determine segments that can produce greater insight. How do "Gold" users behave differently than "Silver" users? How does my conversion look like month to month by cohort? Etc."
Simo Ahava from Simoahava.com
"When talking about data and actionable insights, I believe there to be only one way to consistently deliver on those fronts: by breaking down silos in the organization. You can have the best analytics partner in the world, you can have an experienced, expert developer / IT team, you can have marketers worthy of their own TV show, but if there is a lack of communication between these stakeholders all your efforts can be in vain. The journey from data to business growth starts with a healthy organization, founded on communication and not confrontation, constantly inspired, motivated, and curious about data and the possibilities it has across the entire organization. So my number one strategy is to find the pressure points in the organization and treat those before even mentioning a single three-letter acronym such as KPI, KBO, or CLV."
Stéphane Hamel from Immeria Consulting Services
"I often get asked “how do I start?” - well, with experience comes some wisdom… I hate to reinvent the wheel and at some point in my career I stumbled upon the Six Sigma concept of DMAIC - Define-Measure-Analyze-Improve-Control. The whole methodology covers many things and offers a ton of useful data-driven tools, but just consider this:
- Define the problem or hypothesis, stakeholders and scope of analysis - you know, that whole “if you can’t measure it, you can’t manage it” (attributed to W.Edwards Deming) or “how you measure success depends on how you define success” (coined by Jim Sterne) - nice statements I rarely see clearly articulated into specific action steps;
- Measure and gather relevant data and conduct basic analysis to spot anomalies;
- Analyze correlations and patterns, put your statistics and visualization skills to work;
- Provide insight and articulate improvement options, and finally;
- Control the change by keeping an eye on relevant metrics and KPIs.
Simple enough isn’t it? Now try it! This is how you turn data into insight."
Tim Wilson from Analytics Demystified
"Recognize that the two main ways analytics can drive value are performance measurement and hypothesis validation, and those are two fundamentally different things.
Performance measurement is where dashboards live, and it’s all about the past. It answers the question: “Where are we today… relative to where we expected to be today at some point in the past?” It should be KPI-driven, objective, automated, clear, and concise.
Hypothesis validation, on the other hand, is about the future. Nothing should ever be labeled as an “analysis” that doesn’t have a clearly articulated hypothesis. And, that hypothesis should be qualified to ensure that it has the potential to drive action. I like to capture hypotheses by completing two fill-in-the-blank statements: 1) “I believe __________.” (this is the hypothesis), and 2) “If I am right, we will __________.” (this is the qualification). Having the discipline to get these statements written down and refined saves a lot of time spent wandering through the data and producing charts that are the dreaded “interesting, but not actionable.”
Todd Belcher from BlueConic
"For me, the number one strategy involves connecting data to systems (technology) or processes (people) where action can be taken or automated. This implies that some forethought has gone into customer data and marketing automation as a whole, answering questions like "What data points for known or anonymous users are indicative of interests or intents?” and "Where can we actually make use of this data?”. If the current marketing technology stack doesn’t enable any productive response to the second question for any reason, it’s time to reconsider the components of the stack."
Tom van den Berg from Online Dialogue
"Just data is not enough. Data is valuable only if it helps a company make better decisions. Data should tell you something. If people, who didn't saw the data before, see the data they should have the feeling: I need to take action now. Many companies are collecting and reporting on a lot of data, but are not using it to change something.
My number one strategy is: "Cut the noise and focus on the most interesting topic / highlights." If you are reporting on data as a web analyst, focus on the most important insight you have and elaborate on this. If you explain all the graphs from a dashboard, nobody will remember. Highlight 1-3 important findings helps to focus and people will remember this."
Yehoshua Coren from Analytics Ninja
"Turning data into actionable insights requires asking a good question. Lots of good questions. Data itself is completely passive. It doesn't "tell you" anything until you ask a question and then seek the answer in your data. I recommend asking questions that focus on uncovering the behavioral intent of users. Formulate a hypothesis. IF I know that a user is interested in X or trying to do Y, THEN I will take such and such action. Having an direction that you are heading in advance of your digging through your data will lead to analysis that gleans actionable insights. Of course, you'll need to be careful of bias. Remember that the answer to your question may not be something that you expected. Be open and curious. But be clear about whether you'll be making a UI change, marketing campaign adjustment, change to copy, running an a/b test, etc as a result of the question. Bottom line, asking good questions is the key."
Zorin Radovančević from Escape Studio
"I strongly believe only a clear internal capability improvement strategy applied to all organizational levels will yield in an ability to truly harness data and generate any kind of a competitive advantage.
Start as soon as you can. Iterate in small increments. Build the best multidisciplinary team available which should be truly savvy in the ways of your business as the entire focus should be on predicting business outcomes. Use internal IT or outsource to build a set of tools which represent a collaborative, scalable and simple to use decision support platform. Give your team free and transparent access to data. 'Force' them to produce simple insights as soon as possible regardless of the predicted outcome and let the modelling games begin."
You have just read 7.000 words of incredibly useful advice that you can directly apply to your business!
I hope you have enjoyed reading these awesome tips and strategies as much as I do! So that you can improve your skills to turn data into actionable insights.
Now it's your turn. What do you think about data and actionable insights? A comment or share is very much appreciated!