October 6, 2016

View Post

Businesses, large and small, are increasingly turning to Business Intelligence (BI) and Predictive Analytics to help manage operations and improve company performance. These two disciplines are closely related and often intertwine, but there are some key differences as well.

Let’s briefly illustrate each of these in the context of a common business application. A BI analyst’s job could be to produce a financial dashboard for the CFO identifying the company’s year-to-date performance on several key metrics, such as sales revenue. To do so, the analyst may leverage a BI software platform to help collect, organize, and transform data, as well as create visualizations to display these insights.

On the other hand, that same company may turn to a data scientist using predictive analytics to forecast what the sales revenue will look like for the remainder of the year based on patterns and trends in existing data, plus identify what factors impact that revenue (and how). This, too, may require the use of a software platform, especially to execute the complicated mathematical and machine learning algorithms that predictive analytics often involves, as well as to coalesce and share the findings back with stakeholders.

At a high level, Business Intelligence and Predictive Analytics have common ground in analyzing data to deliver insights that inform and influence company decisions. Oftentimes, even the software utilized overlaps, as both require data intake and visualization.

The fundamental difference, is that Business Intelligence aims to answer “What happened?” and “What’s happening now?”, whereas, Predictive Analytics attempts to answer “What will happen?”. In other words, BI can tell you past events along with a post-mortem analysis of the “Why” aspect of those events, while Predictive Analytics looks ahead to future events and trends. Simply put, BI is reactive and Predictive Analytics is proactive.

Ultimately, it’s a combination of both Business Intelligence and Predictive Analytics that gives analysts and executives a full view of past, present, and future. This allows companies to both understand what has happened AND take actions toward a better future, setting themselves up for success in a competitive market.

Are you ready to start leveraging the power of Business Intelligence and Predictive Analytics to achieve your business goals? We can help!


August 9, 2016

View Post

It could be said that in every process there is another equal and defining process. Each process then evolves, creating more chutes and ladders than can be found in the infamous children’s board game.

To help you navigate this and come to your single truth in Domo, we have scaled the steps to a successful project for getting started with Domo’s business optimization platform.

Here you may recognize four of the primary principles of analytics with one very important step at center: Clean. It will be the single, most important task in preparation, but flows into each step as we begin to look at them and why they are important.

Step One: Plan

In order to see power and immediate value from Domo, your first priority should be to establish a singular objective. While each of the departments in your company may want different things, it’s more effective to start with one objective you can focus on in the learning phase to get the most from your Domo experience.

By setting out and defining key questions that support that objective, and then by identifying the metrics that will align to answer those questions, you are creating a path to success founded on what matters most to you and your business. You have now created the parameters that will cause actions and decisions -- not just stats and pretty bar charts.

Step Two: Collect

You have planned and put into motion the insights that will drive influence for all team members to create change. Now let’s figure out where your sources of data reside and how you are going to use them. If you know where your data is, you and your project team can plan a strategy for all the ways it will connect to Domo. Locate your sources, outline them, and have a clear path of how you connect, schedule, and retrieve your data.

Some sources have dimensions, objects, and drill downs that can get convoluted quickly. Keeping your plan in mind, pick only those data sources that support your singular objective for this exercise. Obviously all companies have many sources that overlap between departments that will make it easier as you roll out to each team. For now, focus on the stepping stone in front of you. You will build a path.

Step Three: Clean

This is where that data governance comes in. You know what you want to see, and you know where to go to find it. How does it speak to your objective? With those dimensions and objects, are you going to need to render any additional layers or push normalization prior to your Domo import?

One of our recommended best practices is to clean up any data on the back-end once you bring it into the tool. Your data specialist will work with the project team to define every aspect of this. A strong plan will support this, and keeping a solid objective in mind will help you to avoid what's known as "data dump syndrome." Again, you are focusing on what matters for this exercise.

Step Four: Analyze

You guessed it -- it's almost time for the visuals. Digging in to get a tidy data set has ensured your analysis is error free. Recognize there could be some data quality issues that will need to be corrected through validation. That’s the best part: you may find things not previously visible to you, and now Domo gives you the insight you need to proactively address any data quality issues. Now is the time to keep those key questions in front of you at all times as you plan the visuals to analyze. Remember to compress stories into a very concise message.

Your project team is going to encourage and support that "60 Second Story." What can you tell your Executive Stakeholders in 60 seconds that captures the snapshot of their question and provides an answer? With drill down paths and other supporting artifact, you will begin to create an ever moving cycle of engagement.

Step Five: Reporting

Visuals are the ribbon at the finish line. More than just a pretty picture, hopefully the considerations have set you up for more than just a bar chart or heat map. You have set in motion the the stories that compel action. Clean data, concise design, targeted to the specific audience you are trying to answer questions for.

The visual perception should quickly and effectively communicate the meaning. Your project team will look to suggest the best visuals that support that. Through the Domo training opportunities such as "Help Center," "Example Cards" and "Dojo," there are numerous resources for showcasing each card type and what stories it best displays.

Other things to consider:

  • Will I need to Compare or Measure?
  • Single Value or multiples that have no order or must follow time structures?
  • Am I looking for Correlations?

This breaks down things like Nominal, Ranking or Correlation charts. This can be represented in Domo’s bar charts, graphs or even scatter plots.

The Domo help section has tremendous resources when you are ready to build cards for deciding. We at Big Squid are excited to help you in your process, and can help get you to that finish line with Gold Stars and the highest value from your Domo engagement.


June 23, 2016

View Post

Do you remember the very first web application you ever used? Not software, which may have been a videogame or productivity application. I’m talking the very first time you went to a dynamic website and interacted with it?

If you’re not very old, it might have been Facebook, Instagram, or Snapchat. If you’re a bit older, that might have been posting on Slashdot or using AOL instant messenger.

The further back you go, the less it felt like the rich web and mobile applications we use today. But even way back in the AOL days, they all had one thing in common: the developers building those apps were building them around four very simple operations.

Create, Read, Update, Delete

Modern frameworks, often just reference these operations as C.R.U.D. pronounced like a Puritan swear: “Ah, crud!”

The beauty of these four operations is that they explain almost everything one could ever want to do with data on the web and almost everything one could ever want to do with data, period. At least they did.

That’s all changed.


May 10, 2016

View Post

We often hear about examples of how organizations are able to transform using insights from their data, but individuals can also benefit from data collection and analytics in their lives.

For individuals, transformative behavior change often starts with a bit of introspection followed by a resolve to make a change, and then crystallizes (or fails) with the help of our brain's habit circuitry. (Charles Duhigg makes a great case for this in his book The Power of Habit).

People have long been collecting and monitoring qualitative and quantitative data points in their lives to fuel and guide such change -- think of journals, hour tallies, and budgets. Often taking a step back to take stock of our 'historical data' and where we'd like to go allows us to hone in on areas that require attention and adjustment, which might not have been obvious to us during the whirlwind of our daily activities.

More recently, such data collection has become much more efficient and automatic, allowing us to more clearly see trends and identify behaviors without expending too much energy in collecting and visualizing that data.

This is especially true in the following realms:

  1. Personal finance
  2. Fitness & health
  3. Productivity

For example, financial tools like Mint are enabling users to track and monitor spending behavior. Fitness trackers like Fitbit are enabling wearers to make sure they're moving and exercising enough during the day -- valuable as we spend more and more time sitting in front of screens.

Similarly, tools like Google Calendar and Toggl are allowing us to track time spent on deliverables, building critical skills, and other high-value work. The simple act of adding up hours can provide insights into whether we are spending time on the most important tasks or getting sucked into low-value tasks.

If used the right way, these tools capture and provide the data we need to help us shape and adapt our behavior more quickly and intelligently.

What are the data-driven habits in your life? Which tools are you using to drive personal transformations?


May 2, 2016

View Post

Cross Channel Attribution has been the “talk” for quite a while now, but even today, it's still an area where many marketers are reluctant to do more than just dip their toes in the water, which is understandable given the complexities of it.

I believe the main reason for the hesitation is that there's no cut-and-dry way to go about it. There are so many articles out there on the subject, and they certainly don’t always align. And depending on who's making the recommendation, there's typically some level of bias involved.

Whether you're managing all of your media channels in-house or using multiple vendors, each vendor or internal team is likely going to go with an approach that, at the end of the day, will “prove” that the channel they are managing deserves the most credit.

So who do you trust? Which way is the “right” way? And if you try one out, are you even guaranteed it will be better than what you are doing now? Maybe. Maybe not.

Regardless of whether or not you choose to make a change to how you're currently attributing transactions across your marketing channels, there are still some straightforward -- and very important -- things you can do to set yourself up for success. I’ve outlined them below:

1. Implement Proper Tracking

This is probably the single most important thing you can do. If you don’t have proper tracking in place, you'll never be able to truly optimize your media and drive success for your business. When you take cross-channel attribution out of the equation, improper tracking has a major impact on each media channel individually. As a marketer, if it looks like you are only garnering a 0.5 ROI on a Paid Search campaign, you may pause “underperforming” keywords, pull the budget from that campaign or pause it altogether. Suddenly, you see that overall revenue for the next week took a much bigger hit than anticipated, and there were two main reasons for this. First, you're probably not looking at the downstream impact those keywords have on generating transactions that other media channels are picking up (assuming last-click attribution). Secondly, because your tracking wasn't set up properly, you're “optimizing” half blind.

2. Get your Data Into One Platform

Once your tracking is in place, getting your data into one platform should be goal number two. I’m certainly not saying you should necessarily trust one digital agency to actually manage all of your channels for you, but I am saying that you should make sure that the data from all the media channels you are managing is visible to you in one platform. You need to be able to tie all of the pieces together, and it's virtually impossible to do so if your data is housed in multiple places, at least not without a ton of manual work.

3. Collaborate/Plan among different marketing departments

One of the biggest problems marketers have is changing the mentality or structure of their media and marketing departments. Oftentimes, each channel is being managed in a silo, and each department has its own goal to hit for that quarter or that year. They may occasionally talk to each other about some “big picture” items, but there are very few companies that have a fully- integrated department in which everyone works together to achieve a common goal, and where each area of marketing isn't constantly having to make a case as to why they should get more budget than another program should. This is the only way you will ever truly realize your potential, and while it’s not always easy to change over to this mentality, it’s a lot easier to do when all of the marketing and media buying is happening in-house. When you’re utilizing several companies to help do this for you, it becomes much more difficult. But, if you have the proper tracking in place and have full visibility into all of the marketing channels, you will be much more informed to allocate budgets more effectively, and you'll be able to make better decisions on your overall marketing strategy.

4. Test, Test, Test

If you’ve done all of this, then you're in a great place to start testing out different attribution strategies and models. You can now see how everything ties together to lead to the ultimate goal you are trying to achieve, and this will at least help inform you as to what makes sense to test. One model may work great for a retail company, but not for a financial company. There are so many other factors out there that impact your bottom line, and it is certainly not a “one size fits all” world we live in when it comes to attribution. Testing out various models will give you the final piece you need in order to hit the ultimate goal of fully optimizing your marketing budget to get the biggest bang for your buck.


April 26, 2016

View Post

"Information is the oil of the 21st century, and analytics is the combustion engine."

- Peter Sondergaard, SVP, Gartner Research

There’s a great little mom-and-pop ice cream shop where I grew up. An old Korean couple runs the place, and every time I stop by on a nostalgia visit, the same smiling grandma is there asking how I’m doing and knowing what I’m going to order by heart. This is the kind of old fashioned place where they do everything by hand, right down to jotting down their sales in a hand written notebook at the end of the day.

Recently, I asked the owners what they actually do with that data, and I was pleasantly surprised by their response. Sure, they look at monthly sales, inventory numbers, and that sort of thing.

But they had also looked deeper and identified trends that affect their business. When it’s raining, customers tend to come in less frequently. Starting in May, sales tend to shoot way up (kids are out on vacation… and Georgia summers are no joke). The ice creams that are prominently featured in the display case tend to be sampled and then bought the most.

While these conclusions intuitively make sense, the coolest part is that they were actually able to put some numbers behind these insights to influence decisions on staffing and product placement.

Although they would never have called it such, the old Korean couple with the ice cream shop had been practicing predictive analytics.

What is Predictive Analytics?

At its core, predictive analytics is about identifying and quantifying patterns in data, and then using those patterns to make predictions about future events.

For example, if sales are consistently low during the winter months, there’s a pretty good chance that they’ll be low again next winter. If customers are canceling subscriptions after several complaints with customer service, a customer with a similar experience is also at risk to cancel.

Sometimes these patterns are easy to discern; yet more often than not, there are so many variables at play that it quickly becomes impossible to judge on_ feel _alone. For example, consider how many factors contribute to the value of a home -- the size and features of the property itself, nearby schools, the crime rate in the surrounding area, and more. Through a variety of statistical methods, predictive analytics attempts to use available data to scientifically quantify these patterns.

Of course, the fun part is applying these insights. Having a strong predictive model for key metrics is like a real life cheat code. Imagine how empowering it is for ANY business to be able to make confident projections about future trends.

A business can plan for the future and potential trajectory, identify risk factors and opportunities for growth, and consider the effects of different scenarios on key metrics. If that sounds too good to be true, well, there is one catch: predictive analytics can only find patterns in the data to the extent that there are patterns to find.

At the same time, that very principle gives assurance that this isn’t some magic 8 ball solution – these are data-driven insights born from the rigor of statistical analysis.

Predictive analytics aims to take the guesswork out of predicting, and allows businesses to move from being reactive to proactive.

And that’s something that everyone, from Fortune 500 CEOs to sundae-slingin’ grandmas, can appreciate.


April 25, 2016

View Post

We see so much information written about "Big Data" and all the complexity that goes into it. Reading countless articles on the subject can cause someone's eyes to glaze over, and we sometimes lose touch of the idea that all this data can be used not just informatively, but also in fun ways that affect our everyday lives. Regardless of what you may think, it's not just for data companies to compile boring data.

Here are 4 interestingly fun ways big data has become a part of things we use all the time:

Sports:

Ski resorts have been using RFID (Radio Frenquency Identification) for a while now instead of the traditional ski lift tickets. Using the RFID, ski resorts can track which lifts are seeing the most activity, and they can use this data to improve user experience by making changes to the lift process by speeding up wait times. Data can also be tracked to show a specific skiing activity, which can also be viewed on a smartphone, making it possible for metrics like total runs for the day, time on the slopes, and overall skiing progress to be analyzed. Pretty cool stuff for the avid or even new skier.

Television:

Understanding what consumers want is an important part to any industry, but possibly even more so when it comes to television. People aren't just watching cable -- they're online streaming on-demand sources like Netflix and Hulu, and just about every network has some form of these services. All the shows watched on your PC, SmartTV or Smart Device are tracked, and this data is used to offer content that's specific to the user, all in an effort to deliver a better experience for the show you want to see. It is, of course, also used to deliver ads for content on related purchases you might be interested in. Not everything is totally "free" right?

Food:

Food is a passion, and all you foodies and enthusiasts out there might find this interesting. IBM has introduced what they call "computational creativity." This technology can analyze thousands of recipes and understand things like dish composition and ingredient pairings to come up with new recipes that chefs can try. In an age where unique food ideas can be hard to come by, data has found a way to help create new ideas that have never been thought of before.

Travel:

I'm sure everyone who has traveled across the country or globe has had to rent a car at some point in time. With so many rental car companies to choose from, keeping both customer service and product quality at a highest level is very important for repeat business. Well, Hertz gathers and analyzes information using an IBM Content Analytics sofware. What this software does is quickly analyze thousands and thousands of customer comments that are submitted via email or phone. This information, whether good or bad, is helpful, and Hertz is able to use this data to make improvements such as creating faster rental and return processes. Data is also used to show which customers request call backs from Managers to help resolve any issues during the rental process.

So while some people may not be "into" Big Data, it's important to realize that it's all around us, and it helps to improve the things in our everyday lives that we actually are "into."


March 10, 2016

View Post

Cross Channel Attribution has been the “talk” for quite a while now, but even today, it's still an area where many marketers are reluctant to do more than just dip their toes in the water, which is understandable given the complexities of it.

I believe the main reason for the hesitation is that there's no cut-and-dry way to go about it. There are so many articles out there on the subject, and they certainly don’t always align. And depending on who's making the recommendation, there's typically some level of bias involved.

Whether you're managing all of your media channels in-house or using multiple vendors, each vendor or internal team is likely going to go with an approach that, at the end of the day, will “prove” that the channel they are managing deserves the most credit.

So who do you trust? Which way is the “right” way? And if you try one out, are you even guaranteed it will be better than what you are doing now? Maybe. Maybe not.

Regardless of whether or not you choose to make a change to how you're currently attributing transactions across your marketing channels, there are still some straightforward -- and very important -- things you can do to set yourself up for success. I’ve outlined them below:

1. IMPLEMENT PROPER TRACKING

This is probably the single most important thing you can do. If you don’t have proper tracking in place, you'll never be able to truly optimize your media and drive success for your business. When you take cross-channel attribution out of the equation, improper tracking has a major impact on each media channel individually. As a marketer, if it looks like you are only garnering a 0.5 ROI on a Paid Search campaign, you may pause “underperforming” keywords, pull the budget from that campaign or pause it altogether. Suddenly, you see that overall revenue for the next week took a much bigger hit than anticipated, and there were two main reasons for this. First, you're probably not looking at the downstream impact those keywords have on generating transactions that other media channels are picking up (assuming last-click attribution). Secondly, because your tracking wasn't set up properly, you're “optimizing” half blind.

2. GET YOUR DATA INTO ONE PLATFORM

Once your tracking is in place, getting your data into one platform should be goal number two. I’m certainly not saying you should necessarily trust one digital agency to actually manage all of your channels for you, but I am saying that you should make sure that the data from all the media channels you are managing is visible to you in one platform. You need to be able to tie all of the pieces together, and it's virtually impossible to do so if your data is housed in multiple places, at least not without a ton of manual work.

3. COLLABORATE/PLAN AMONG DIFFERENT MARKETING DEPARTMENTS

One of the biggest problems marketers have is changing the mentality or structure of their media and marketing departments. Oftentimes, each channel is being managed in a silo, and each department has its own goal to hit for that quarter or that year. They may occasionally talk to each other about some “big picture” items, but there are very few companies that have a fully- integrated department in which everyone works together to achieve a common goal, and where each area of marketing isn't constantly having to make a case as to why they should get more budget than another program should. This is the only way you will ever truly realize your potential, and while it’s not always easy to change over to this mentality, it’s a lot easier to do when all of the marketing and media buying is happening in-house. When you’re utilizing several companies to help do this for you, it becomes much more difficult. But, if you have the proper tracking in place and have full visibility into all of the marketing channels, you will be much more informed to allocate budgets more effectively, and you'll be able to make better decisions on your overall marketing strategy.

4. TEST, TEST, TEST

If you’ve done all of this, then you're in a great place to start testing out different attribution strategies and models. You can now see how everything ties together to lead to the ultimate goal you are trying to achieve, and this will at least help inform you as to what makes sense to test. One model may work great for a retail company, but not for a financial company. There are so many other factors out there that impact your bottom line, and it is certainly not a “one size fits all” world we live in when it comes to attribution. Testing out various models will give you the final piece you need in order to hit the ultimate goal of fully optimizing your marketing budget to get the biggest bang for your buck.


March 2, 2016

View Post

As many companies attempt to take advantage of the availability of data, they tend to run into one age-old problem: culture.

While formal edicts and new software may be needed to make use of business intelligence, we have to remember that culture is essentially defined as how things are done. Therefore, you have to address culture if you want to do things differently.

Here are five tips that may help make the process a bit easier:

1. Use a process that's data driven

It's important to take some time to gather information on why this change is needed. Gather data on how this change will affect the organization as a whole, as well as all the levels and groups in the organization. How will it help them? What will they need to do, and what challenges might they face?

2. Present your findings back to the groups

Format the data in a way that will allow each group to appreciate it. Remember that you'll be presenting to everyone from executives to the front lines, so your message is going to have to resonate across several different groups. Once presented, give the groups time to analyze, discuss, and provide feedback on the data.

3. Avoid Change Fatigue

Too many changes happening at the same time without prioritization will cause the inevitable rolling of the eyes. Make sure you're not presenting this culture change while another major change is occurring and competing for your organization’s resources. It's important to take the necessary time to properly design the change process, and it's even more important to take time to properly communicate the change to those within your organization.

4. Communication vs. Engagement

Both are critical, but you need to understand the difference between the two. Communication is information, while engagement is involvement. Too often we communicate the change, but we don’t engage those who are going to be affected or are needed to make it happen. Have your people be a part of the change, instead of just informing them that it's happening.

5. Manage the formal and the Informal

Lean on existing strengths within the organization. Get buy in from the influencers within your organization, and remember that those people don’t always have fancy titles, yet they are the ones who people trust and follow. Ensure coherence by avoiding mixed messages, especially from executives and management. Get an understanding of the “Human Side," and deal with actual feelings by having a realistic understanding of the organization’s history, readiness, and capacity.

----

If you're attempting to direct your organization into a culture that's driven by data, it's important to keep these five tips in mind to ensure a seamless transition.

Get your organization on board with the change by effectively communicating why it's necessary, and how it's going to directly impact the company and its employees.

Always remember that communication is key -- even when it comes to data and numbers.


February 16, 2016

View Post

Valentine’s Day wouldn't be complete without a plethora of hectic online searches to find the nearest store that's selling flowers and chocolates, all while praying they still have a decent selection left. It’s a situation that most people will find themselves in at some point, and the results can be fairly entertaining.

Starting off with a light-hearted approach, the real estate company Estately set off on a mission to find out which type of Valentine's Day gifts people in each state Google more frequently than people from other states.

To get the results, Estately ran hundreds of Valentine’s Day-based queries through Google Trends, and then they labeled the states according to what each state searched for the most.