Success in building a predictive model often hinges on your data. A central element of the data concerns how it, and the resulting prediction, is framed around the business challenge - the outcome you’re interested in forecasting. As we apply machine learning to empirically identify the the drivers of outcomes in our historical data, our findings are often governed by the applicability to current data. Is the past relevant to today? Is that data representative?
We’ve witnessed our customers apply several strategies to create robust training data sets (i.e historical data) and therefore maximize their live experience applying these models to real-world business challenges. Let’s explore a few of those tactics and key questions in approaching predictive analytics.
Think about your prediction problem today. What do you know?
Thinking about what you know today is a good proxy for what you need to capture in your historical dataset. If attributes of your data existed historically but are no longer part of your data collection, it likely shouldn’t be used.
When do you know it? / Is your data timestamped?
This is among one of the more critical considerations in preparing data for a predictive problem so we can better avoid look-ahead bias. We are likely to have observed data at different points, with different timestamps, historically.
Framing a forecast around what was known at that time is key, creating a true point-in-time dataset. It reflects that data available to a decision maker at that exact point in time. As we train a model, a machine takes the place of this hypothetical decision maker and learns the set of attributes that led to positive outcomes.
Framing the problem in this way, gives us a reflective representation of the data if we were placed in that period of time historically. As such, it helps mimic our current environment and makes our historical learnings more robust and applicable to today.
As an example, if we’re building a customer retention model, knowing our customer’s attributes as of 12/31/2014 and that they canceled on 6/30/2015, we can’t use data (fees, payments, service calls, offers, etc.) that occurred from 1/1/2015 – 6/30/2015. We wouldn’t have known those data points at the point in time we were interested in making a decision or forecast (12/31/2014).
This extends to attributes that aggregate data over time such as averages (it should only aggregate to that point in time) as well as class variables. For the latter, whether or not the customer is receiving our monthly newsletter might be an attribute to help predict retention. But, unless we know when they started to receive it, it’s likely to yield inaccurate results potentially due to look-ahead bias. If they started to receive it on 3/31/2015, and we didn’t timestamp that event, our use of this variable would assume that they were receiving our newsletter on 12/31/2014. Because they canceled, naïve algorithms might assume a negative relationship between newsletter subscriber and customer turnover.
Rest assured, many companies applying predictive analytics are grappling with similar issues. In the near term, you may identify data points that would be useful for predictive modeling, but without proper time context, can’t be used.
It may naturally prompt better data collection initiatives, and this is positive—predictive models are a perpetual work-in-progress, and there’s no bigger input to quality than the data underlying a model.
Determine at what point you could take action from a forecast.
Do some inputs compile at month-end? Are different data points available at different times? Data is frequently known, available, and therefore actionable at various points in time. Think through the optimal point to “strike” your data and take action today. Knowing that, construct your historical training data in the same manner. Waiting longer for key data points delays our ability to act; acting earlier without key data points can be just as perilous.
Also think about the window of time into which your forecast is set. Can you take the same actions on a forecast over the next week versus one that is expected to occur over the next year? If you can’t mobilize within a week, then that prediction timeframe is of no use since you can’t act on it.
**Consider extending out the forecast horizon to be more attuned to the
base churn rate.**
Use the historical base rate as a guide. If you build an employee churn model with a forecast horizon of 1 day (i.e. who’s likely to resign tomorrow?) when your annual turnover rate is <10% (ie the base rate), the forecast horizon and natural tendency of churn are somewhat incompatible. Turnover is so infrequently observed on a daily basis that identifying its drivers becomes difficult. The model is unlikely to have meaningful levels of accuracy, as a result.
Framing data around the outcome you’re hoping to model is a key component to building a predictive forecast. When constructed properly, it allows you to understand what drives those outcomes and therefore areas to focus on to positively alter those outcomes. The tactics and strategies outlined will go a long way to ensuring success in this endeavor. If you are interested in learning more or if you have any questions, please feel free to contact us today at Big Squid.
“Don’t Make Me Think”
Those familiar with user experience design in the web application development context, might have heard the phrase “don’t make me think” coined by Steve Krug. This framework calls for usability and simplicity to be at the forefront of web design, even when developing the most complex of systems. Think about the last time you submitted a form online, perhaps while signing up for a new account or ordering a new pair of shoes. You most likely did not have to go read a manual to understand how to use the website - it was obvious even with minimal instructions.
The same concepts apply when building your business intelligence solutions. Most tools today allow us to create visualizations very quickly and iteratively, however there is still some design thinking that must come into play in order to build effective, actionable, and adoptable data visualizations.
Know Your Audience
At a high-level, effective data visualization, also known as card building in Domo, starts with understanding the audience for the dashboard(s) on which the visualization will live and the business questions they are looking to answer.
Customers often start dashboard design by identifying a function or department they want to onboard onto their new BI tool. The general goal might be something like “understanding how effective our marketing efforts are.” If we take this one step further, we’ll learn that one dashboard might not be an effective way to achieve this business goal. The executive audience for this goal is interested gauging the overall effectiveness of marketing efforts in the larger company strategy but the marketing managers would be better served by a more tactical dashboard in which they can make decisions on specific campaigns.
Understanding the audience and the story each unique audience is interested in will help guide the level of granularity needed in the visualization.
Choose Chart Types Wisely
As you start to understand your audience and the questions they are looking to answer, start crafting card designs that will answer those questions as a story. Your data visualizations will typically fall into 2 categories: actionable insight & context card. The actionable cards are the ones that tell you how much progress you’ve made towards a goal this week, how many opportunities you need to close to meet sales quota this month, which customers need some outreach this week, etc. The context cards supplement those cards, showing the impact of various influencers or dimensions on key metrics, such as a view over a larger period of time, a map of where the business operates, etc.
Use the right chart type to tell the right story. To name a few examples:
Line charts are good for showing trends over time and continuity. Don’t use a line chart if your x-axis represents something other than a date or time period.
Bar charts are good for showing comparisons between different categories. Use a stacked or grouped bar to show additional comparison between related metrics or another dimension.
Donut and Pie charts are good for showing composition or part-to-whole comparisons.
Line + Bar charts are good for showing trends over time plus the influencing metrics. For example, use the line to represent an occupancy rate overtime and the bars to show occupied & available units.
Map cards are good for showing geographic distributions.
A Few Design Tactics
If you’re a new card builder, it is tempting to build quickly. Do so, but don’t forget the low-hanging fruit that will make your visualizations and the overall dashboard much more consumable. Here are a few tactics to get you started:
1. The bar charts are a very popular type of chart, and in many cases justifiably so. Think about colors in the context of the dashboard, not just the current card. You might find that you initially have a string of blue bar charts across the same dashboard. While this might tell the necessary story, use color to help the user quickly identify each card. For example, maybe you make your cost-related card red, your profit-related card green, and the application count blue.
2. Design for adoption. If it’s not easy to understand, it will not be used. Again, “don’t make me think” to understand the point of the visualization. The visualization should be designed to help me ask questions to improve & make decisions on my business. If this card ended up in an online article or newspaper, would the average readership understand what it’s trying to say? Are there labels, hover text, or descriptions that could help make it more consumable? Are the numbers formatting correctly as currency, percentages, etc.?
Iterate but Keep It Simple
Once you’ve drafted a card, you can and should continue to iterate over it based on feedback from the business users. As feedback comes in, think before incorporating it into the card, so as to not over complicate the card. Would the new ask work better as a separate card or as a drill path? Will adding an additional metrics make the visualization confusing or will it help provide needed context?
Hopefully, this guide has helped you understand card building a little bit better, but if you need any assistance, please feel free to reach out to the Big Squid team.
In recent years, organizations of all sizes are looking for more effective ways to proactively identify and monitor business trends and adapt accordingly in a timely manner. This is the heart and soul of Business Intelligence or BI.
With the right BI solution in place, you can improve decision making and management at all levels of an organization, gain valuable insights into consumer behavior, turn those insights into actions, and improve overall efficiency.
In case you haven’t already discovered, there are many business intelligence tools available on the market, but what do you REALLY need in a high quality and meaningful BI solution for your organization?
Perhaps, your organization is already using a particular tool, but what’s missing that’s preventing you from extracting crucial information for success?
Today, we’ll answer this question in the context of the Domo platform, which is one of the top BI and analytics tools available.
1. **A way to collaborate across your organization**
A good BI solution makes it easy for everyone from top executives to smaller teams to access, analyze, and update information anytime, anywhere, enabling better-informed and timely decisions across the board.
The Domo Platform offers a way to collaborate directly with your peers and co- workers via Buzz. Domo Buzz is highly effective through the mobile application, as well as the desktop version. Users can discuss performance, answer questions, and collaborate on solutions or strategies based off of metrics right at their fingertips.
Buzz is also offered to all individuals in your organization, whether they have full access to Domo or not, so no one is left out of the conversation.
2. **Customized solutions for every user to analyze**
To glean any value from your BI solution, it should be simple for you to filter out the data that is irrelevant, and only see what metrics really matter to you.
Through security configuration options and recommendations based off of best practices, take your everyday metrics and apply filters, so each user can have a tailored platform and set of visualizations upon login to Domo.
No longer will your employees be inundated with data that they do not care about or is of lesser value. For example, your managers overseeing the Northwest division can log in and instantly see sales metrics for the Northwest division only, allowing them to navigate the platform and information with ease.
3. Get alerted when a key metric changes and needs your attention
The best business intelligence solutions deliver real-time data, since the key to BI is making decisions based on the most current data, without which, you’re at a huge disadvantage. Therefore, it’s vital to know of any changes as soon as they occur.
Leveraging alerts can enable all users to set alerts based off of key metric standards. Using Domo, you’ll never miss an important change in your performance with alerts on your desktop, email, or mobile that prompt you to view your metrics, and take action right away.
4. **Predictive modeling to drive proactive business solutions and conversations**
Predictive analytics allows you to gain a competitive edge, in that you can use real data to identify valuable customers and improve retention rate, pinpoint areas where your competitors are failing, and make accurate predictions about the future of your business.
Through the use of Domo, the Big Squid Predictive Toolkit is available to assist everyday business users understand forecasting in a way that is visually appealing and insightful, without the need for a full fledged statistician on staff.
Ready to jumpstart your business with a smarter business intelligence solution? As a Domo certified partner, Big Squid is ready to help you and your team implement Domo in your organization, and gain insights like never before.
CONTACT US TO Learn More:
If you’re a business owner, having a clear picture of your company’s current state and performance is imperative, but so is knowing where you’re headed. Having this comprehensive understanding is key to determining how to effectively make progressive changes within your organization and gain a competitive advantage.
This is where the field of predictive analytics comes into play. This discipline incorporates advanced analytic techniques such as data mining, statistics, modeling, and machine learning to predict future outcomes based on patterns observed in historical data.
While predictive analytics has been around for quite some time, more and more businesses are now beginning to see and understand just how powerful it can be. The information surrounding predictive analytics is overwhelming, giving rise to all-too-frequent misconceptions and myths.
If you want to use your data wisely, then don't let these myths fool you! For this reason, we’re lending you a hand today by separating fact from fiction and putting your questions to rest.
**Myth #1: A PhD is a prerequisite for implementing predictive
A relic of the past, this is fortunately not required today. You DON’T need a team of data analysts with years of statistical experience and training to put predictive analytics into practice in your organization. Today's tools have made predictive analytics increasingly accessible to business users, empowering them to conduct in-depth analysis and forecasting on their own.
It does, however, require the user to understand the business inside and out, especially its ultimate goals. For example, a marketer should have a solid understanding of consumer behavior in order to successfully take actions toward the desired goal.
Myth #2: The selected algorithm matters most.
While the algorithmic approach that you use is significant (think fancy math), it’s certainly not the most important element. If the resulting model provides little to no value to the business, then why should it carry any weight?
A properly-applied algorithm is a required aspect of building a predictive model, but ultimately, the quality of the data on which that model is built matter most.
For example, consider two companies that build a model to predict when men of a certain age group will purchase a particular product in store. While they may both apply the same algorithm in building their predictive model, if one has captured more relevant attributes about their consumer base, over time, and across a diverse mix of consumers, it’s likely to have a more robust outcome. This translates to greater confidence in basing key business decisions on these predictions, and engaging with future customers.
Myth #3: Predictive analytics is expensive.
You may have heard on more than one occasion that predictive analytics is an expensive venture. As noted earlier, with today's tools, you don’t need to spend money and time on hiring a full-fledged team of expert statisticians and data analysts to execute predictive analytics in your organization. Also, since you can optimize your key business metrics and identify risks, you can save money, spend less, and see significant ROI potential to justify costs. Once you have a better idea of what drives the highest ROI in your business, you can devise and adapt your plan of action accordingly.
Myth #4: Predictions = Prescriptions.
Just because you’ve formulated a set of predictions, it doesn’t mean that your task is complete. The key is to glean insights from those predictions, understand the drivers, and make them prescriptive by conveying actions clearly to stakeholders. Old rules apply—know your audience and deliver these actions in an impactful and consumable manner.
Keep in mind that with high-quality data, you can devise more accurate predictions and models, but you should always keep a close watch on shifts in the drivers of your model or other events that may have an adverse effect on the forecasts' validity.
Are you interested in implementing predictive analytics in your business? Visit us today at Big Squid, and we’ll help you discover the answers to your key business questions, forecast future trends, and turn those insights into actions. We also invite you to give our innovative Predictive Toolkit™ a try, a powerful predictive solution for businesses of all sizes.
With Domo in place, key members of your organization are given the unique ability to properly assess and interpret data to drive real business results. At Big Squid, our team of Domo certified business consultants have assisted hundreds of businesses in implementing Domo successfully, allowing them to discover those actionable insights, and make vital business decisions accordingly.
What does it take to be a Domo business consultant, and how can they help you achieve your particular business goals? Let’s answer these questions by delving into the daily efforts of a business consultant.
First, no two days on the job are the same, but there are a few unifying elements in bringing an engagement into full fruition—strategizing solutions, collaborating with team members, and communicating challenges, risk and opportunities to the client.
The primary goal of each and every engagement is to identify key business questions, and to develop a solution and strategy to uncover business insights that drive action.
In order to do this, a Domo business consultant will begin by helping you hone in on what defines success for you, be it an increase in financial performance, minimizing funding to sources with less ROI, increasing retention rate, etc.
Next, he or she will work with you to determine what you want to focus your efforts on, and to clearly define the complete scope of your project within your budgeted timeframe. The consultant will also discuss the potential roadblocks or obstacles involved, and will make all team members aware, so you, the customer, can make an informed decision about how to proceed. Based on this information, you and your consultant will create action plans for mitigating these obstacles.
With the project guidelines in place, your business consultant will communicate with you on a regular and ongoing basis regarding status, scoping, planning, and delegation, and will also conduct workshops to walk you through the ins and outs of the Domo platform, so you can utilize it to its full advantage. Maintaining this line of communication is essential for keeping the project on track, and achieving your goals by the targeted completion dates.
To be a successful and dependable Domo business consultant, it takes teamwork, organization, and time management. Are you ready to start visualizing your data, and setting your business on the path to success? Let Big Squid’s seasoned business consultants help you with your Domo implementation today and empower your team with the information to transform business insights into actions.
Get Started Now:
Don’t let Domo deceive you. You may think it’s a just a great business intelligence platform where you can not only store your data in the cloud, but bring all that data together to create innovative and meaningful reporting for your business. But, it’s so much more than that—it’s a living, breathing beast, (as you may have literally learned in our previous post on “Beast Modes’), and you have to be willing and ready to take on the fun, exciting, exhilarating task of feeding, maturing, and advancing that beast to a whole new level. Here are a few tips and ideologies to help you do just that.
Is what I see actionable?
Domo offers a host of powerful tools that enable businesses of all sizes to make sense of their data, and translate it into some meaningful results. One of the primary steps in utilizing the Domo platform is building KPI cards, which essentially allow you to begin visualizing your data.
When constructing these cards, one particular Domo feature that will unexpectedly prove to be useful in many different situations, but is often misunderstood is Beast Modes—calculated fields you create, which are implemented in syntax similar to MySQL.
Beast Modes typically fall into two categories:
(1) Calculations applied to every row of a dataset
(2) Aggregations that are done based on the filters applied to the card.
When should you create a Beast Mode?
Here are some of the tasks within Domo that are best suited for Beast Mode.
(1) Calculating dynamic ratios, or aggregations that should change as the underlying data is updated, i.e. Cost-To-Profit Ratio, Overdue Orders Ratio
(2) Simple arithmetic between columns in the datatset, for example, Net Order Value that sums up the Line Item, Shipping, and Tax values for each row
(3) Date functions, such as extracting the day name from a day to display on the card, or calculating the differences between dates. For example, when working with project-level data, you can use the DATEDIFF() and CUR-DATE() functions to calculate how many days a project is overdue.
(4) Creating categories or groupings for your data that don’t already exist. You can categorize orders, projects, etc. as ‘Past Due,’ ‘Due Today,’ or ‘Due in the Future,’ using logic implemented in a CASE-WHEN statement.
(5) Building a complex summary number. Sometimes, you want to display more than one number in the card’s summary. In this instance, you can do this using the CONCAT() function.
As you can see, Beast Modes within Domo can assist in performing a variety of functions, and can make it simpler to interpret your data, and derive actionable insights from it. With a little bit of technical skill, anyone within your organization can learn how to effectively utilize all aspects of Domo to quickly drive impactful change.
You don’t need to be an experienced data whiz to reap the benefits of Domo, but if you could use a helping hand in implementing it in your business, then please visit us today at Big Squid, and our dedicated team of consultants will help you get started.
A robust dashboard can be invaluable to a CEO if it’s created with actionable insights in mind.
Big Squid has presented hundreds of dashboard solutions to CEOs from all kinds of organizations, and yes, there is a general unifying theme to their requirements and key questions—growth, profit, period over period deltas, and company culture.
Let’s examine growth; how does a CEO define it? It is a combination of factors—company size, sales and marketing growth in comparison with a previous period, Marginal Profit over time, forecasts, and employee sentiment over time.
Once you understand how a CEO defines growth, it’s not a complicated task to help him or her determine whether or not he or she is winning or losing in terms of these key growth metrics.
Because different verticals measure sales and revenue differently, we’ll focus on two of the most common scenarios today—software and retail—and we’ll showcase key visualizations for both business models, and the questions that can be answered with these visualizations.
Visualization #1: Sales (ACV) and Marketing Growth (Rolling 13 months, this year vs. last year by month)
Domo is one of the most powerful platforms available for transforming massive amounts of data into tangible business solutions. To achieve this, you can pull data into Domo from many different sources, and visualize it in such a way that you can quickly answer your key business questions.
When it comes to making sense of your data, it’s not always in the cleanest, most digestible format to begin with, and perhaps you want to consolidate data from different systems into one dataset for visualization. Fortunately, Domo provides several solutions for Business and Technical Users to accomplish this—Magic ETL, Data Fusion, and SQL Data Flows.
Let’s explore these robust tools in greater depth, so you can begin leveraging the full power of Domo, and start seeing real results.
(1) Magic ETL
- Business users who understand the context of the problems they are trying to solve, but may not have the SQL or coding skills to transform their data.
- Technical users who want a quick and easy way to transform their data without writing any code
- Magic ETL provides the balance between simplicity and flexibility that makes it a breeze for the user to manipulate their data.
- The click and drag interface allows the user to format their data in a state that is ideal for creating visualizations.
- Some of the more complex transformations may be better suited for another transformation method.
(2) **Data Fusion**
- Business Users and Technical Users who want to quickly combine datasets that don’t require any transformations or cleansing.
- The visual interface provides a quick and intuitive way to combine your data without writing any code.
- Fusions also provide an efficient means of processing large datasets in seconds.
- For data that requires any transformation, such as calculated fields or cleansing, you may want to consider using a different, more applicable method, like Magic ETL.
(3) SQL Data Flows
- Business users who can understand and write SQL transformations
- Technical users who are comfortable with SQL
- SQL Data Flows provide the full flexibility and functionality of the SQL language.
- They also give users two choices—MySQL and Redshift—that can handle all sizes and forms of data.
- As you may suspect, SQL Data Flows requires knowledge of SQL coding to successfully perform any transformations.
The multitude of options available within the Domo platform including the tremendous power of data flows allows users of all different skill sets within your business to gain valuable and virtually limitless insights from your data.
Need help getting started with Domo, and giving your business a boost? Our team of experienced consultants is here to help!
Google Analytics (GA) is an essential tool for any and every business. Not only does it provide valuable insight into the online behavior of your audience, but it also gives you a sneak peek into your best-performing content, allows you to set up goal and conversion tracking and much, much more.
But did you also know you can connect your Google Analytics metrics to Domo?
Domo allows you to pull all of your Google Analytics data into one, integrated dashboard inside your Domo Instance, allowing you to add context to your web analytics and leverage your data in ways you never before thought was possible.
By utilizing the Google Analytics connector within Domo, you can build analytics dashboards that allow you to view and visualize your data however you want, giving you the ability to organize your data in a way that works best for you. You can connect your performance metrics together to gain complete visibility into your campaigns, and identify key trends and behaviors among your visitors. The possibilities are endless.
Convinced yet? If so, follow these three must-haves for connecting your Google Analytics to your Domo instance.
1. The Appstore
Definitely start with Domo’s Appstore, where you’ll find multiple Google Analytics “QuickStart” apps that can be used to power up standard key metrics used by marketers to answer their business questions. These QuickStarts are amazing at producing complete dashboards with just the click of a button. In addition to the QuickStarts, there are other custom apps that measure things like Engagement, Revenue and Product performance. There are currently more than 10 apps just for Google Analytics in the Appstore.
Get connected with Domo’s GA Connector. There you will be able to bring in DataSets around standard GA reports you want to see in Domo. Below is a list of 8 connectors that are current selections within Domo.
- Base Metrics
- Campaign Metrics
- Device Metrics
3. Custom Data Sets
Domo can also set up custom data sets outside of the “Standard” reports from the GA connector. Many times, clients are tracking custom Dimensions, Conversions, Variables, and more inside Google Analytics. Clients can reach out to Domo Support to help get these additional custom data sets created and then use to create visualization and dashboards in Domo.
Google Analytics already provides you with valuable information about your site performance and campaigns, but Domo can provide you with dynamic visualizations that take your GA data to the next level.