Analytics: Turning Data into Management Gold

** This article originally appeared on Applied Marketing Analytics (Volume 4/Number 4/2019) by Henry Stewart Publications **

Competition is the driver behind the growth of B2B and B2C marketing analytics. Management is always looking for ways to beat the competition, and increase the bottom line. There is no shortage of metrics to monitor and track; the end goal being to meet and exceed KPIs while maximizing ROI.

The effectiveness of an organisation’s marketing analytics ecosystem can be directly linked to the actionable insights provided and the actions taken. How deep one digs to get insights is linked to success. The simplicity of how those insights are communicated to management drives decision making and action.

There are lots of vendors out there. To make critical and timely business decisions, management needs analytics platforms that communicate data analytics effectively via dashboards that are comprehensive yet simple to set up. This approach can potentially result in actionable insights. Moreover, such insights can prove to be the key driver to what management needs to promptly make decisive actions that impact the business immediately.

The best way to leverage this powerful tool is a matter of choice.

All forms of analytics approaches follow the mobile, laptop or desktop user journeys. Management should understand the most popular journeys and make business decisions based on the behavior of their users at each step of the journey.

The insights gained from user behavior can be the impetus for decisions that will allow for stronger user retention initiatives.Such initiatives could include rewards based on referrals immediately after users sign up or setting a lower threshold to open an account. One small perk can lead to higher retention rates.

Consider a real-world example. Hershey, one of the largest US snacking companies utilizes analytics to ensure that it has fully staffed sales teams during the busiest holiday times of the year. To ensure it is adequately staffed during busy periods, it uses a system process to predict attrition (REF1).

By utilizing workforce data, Hershey has created a predictive retention model that supports the company in identifying flight risk at both an individual and macro scale. This allows the company to increase staff retention and plan for hiring needs months ahead to ensure it has the right staff on hand when needed most. This successful strategy makes sure that at times of need, the company has the right staff on hand.

INSIGHT-DRIVEN MANAGEMENT: DELIVERED BY THE DASHBOARD

Current marketing analytics platforms employ simple ways to create personalized dashboards for anyone by anyone, whether the marketing team wants actionable insights for campaigns or the chief marketing officer and vice president need to make critical time-sensitive decisions.

Insights pour out of the marketing machine. The critical issue is how to direct those insights to those that can take immediate actions to impact the business.

To harness the power of these insights requires a flexible framework that channels the right insights to the relevant stakeholders. This is the key to realizing the benefits of data analytics — transforming an organisation that was once a jungle of insights into a well-oiled insight factory.

Consider the following example. Management for an online B2B mobile accessory distributor has set up a cart abandonment dashboard for the North America region, so it can identify which areas of the user experience for businesses shopping in the Northwest require optimization.

A cart abandonment team has been set up to receive these insights as soon as they are available. The team sees that users are dropping off at the same spot on their journey from registration to checkout after consuming the same content. The team alerts management to the issues, who make a business decision to drop the content provider and switch to another that can serve up more relevant content based on the user journey. The cost is a bit over budget, but the returns quickly make up for it.

The dashboard makes it possible to see the big picture, highlighting what is most important for each stakeholder, and enabling them to make decisions on the spot.

Top analytics platforms and their dashboard functionality provide different capabilities, which will be elaborated on in next section, and will help to evaluate what platform can meet management needs.

The primary emphasis of dashboards is for reporting purposes. The best dashboards use visual aids such as timelines, heat maps, graphics and charts to communicate information to management. This information ranges from revenue to conversion rates to demographics and engagement metrics, among others.

The attributes to look for in a dashboard-focused solution

Of the myriad solutions out there, many claim to do the same thing to help management reach their goals. For this reason, management must lay out a road map of exactly what they are looking for, as every organisation is different.

At the core of data management in the context of an analytics tool, management must consider a plethora of variables when thinking about their data and an analytics solution:

• Customization: The ability to customize dashboards per user is imperative. Users must be able to access data they want and view said data in a way that makes sense to them.
• Integration/data blending: With data streaming from multiple sources, including internet of things devices, data warehouses, and unstructured social media data, any dashboard must be able to integrate such data, extract information from the data and narrow down that information to display only what is useful.
• Speed: Does the solution under consideration import data before processing, or can it handle queries on the fly? The answer to this question can significantly impact the ability to access data quickly. There may also be requirements regarding the manipulation of data upon import, as well as questions on data capacity limitations.
• Data visualization: This is what gives life to analysis. While all solutions provide this ability to some extent, it is essential to ensure the solution’s visualization aspects are configurable and have the ability to include dynamic, real-time information. The graphical options must be able to truly meet management’s needs and provide the ability to drill down on results. Management must explore whether the solution can provide reliable presentation options such as heatmaps, chart display options and so forth.
• Reporting: The solution must offer a comprehensive tool that makes it possible to share analytics across the organisation, and if requirements so demand, outside of the organisation too. Reporting must be secure and easily customizable by authorized users, not just the IT department.
• Security: Providing secure access is vital to becoming an efficient organisation. In today’s mobile world, users need constant secured access and to be able to see information updated in real time. Management needs to ensure that all devices are supported and data encryption is supported on each. Management leadership needs to be able to set permissions on the fly, that will protect sensitive data so that such data cannot be accessed by unauthorized users.

Asking the hard questions regarding these attributes will find the solution that best meets management’s needs. Asking the hard questions can make the difference between choosing the right solution from the beginning, or wasting budget on multiple solutions until you the right one is eventually found.

According to Econsultancy and IBM, ‘33% of elite marketers say having the right technologies for data collection and analysis is the most useful in understanding customers’ (REF2).

HOW ANALYTICS PLATFORMS TACKLE THE INSIGHT CHALLENGE

Basic analytics

Google Analytics, the father of web analytics is an amazing tool, and is the standard for measuring acquisition — identifying the sources of traffic to a website or app. However, the metrics are mostly soiled, meaning that the tool does little in the way of meshing the entire user journey. By providing only the pieces that make up the journey, it makes it hard to see the entire picture.

For example, if one had 20 users from Australia all online at once, consuming different types of content, performing multiples actions, dropping off, returning, checking out and using different devices, it would be really difficult to extrapolate the complete picture as it is extremely hard or impossible to connect all the dots using Google Analytics.

Google Analytics works for most online businesses. If what is needed is a robust dashboard experience to deliver, for the most part, soiled information, Google Analytics is a great way to start (see Figure 1). Nevertheless, it has its limitations. Most notably, there are simply too many metrics, without any rationale for their inclusion. For marketers on the front line, Google Analytics may be enough. However, for a management planning strategy, that aims to make an impact, deeper and more immediately actionable insights are required.

Figure 1: Mobile commerce dashboard created on Google Analytics, which shows a range of valuable data in one screen: mobile revenues, top handsets, top mobile content, keywords, etc. Source: Created on Google Analytics demo account, using Justin Cutroni’s customization, available here

The difference between using Google Analytics and a behavioral analytics solution can be likened to the difference between using a disconnected approach vs a holistic one: Google Analytics provides discrete metrics, divorced from the context of the user journey, while behavioral analytics stitches those metrics together to provide the full user journey. The basic Google Analytics dashboard shows plenty of metrics, but it is like looking at moments frozen in time — snapshots of analysis that happened this morning, yesterday or last month. The information is static. This kind of static data is disconnected from the other points of the user journey. This is the very definition of soiled data.

The main problem with using a tool like Google Analytics is its lack of ability to identify users. This is essential when it comes to tracking the user’s actions across their journey. Behavioral analytics, on the other hand, identifies users from the beginning of their journey and tracks their actions in a timeline fashion across the web, on your site, competing sites and all their interactions to tell the complete journey that Google Analytics cannot.

Visual/user experience analytics

Popular visual and user experience (UX) analytics platforms such as Hotjar, FullStory and Inspectlet (for websites) or Appsee and UserX (for mobile apps) are terrific for improving the user experience.

These platforms utilize user session recordings (see Figure 2), which offer app marketers insight into exactly how users are using and navigating mobile apps or websites. The platform makes it possible to follow the customer journey all the way through as they consume content, click and checkout and understand which problems they are experiencing.

They also make use of heatmaps (see Figure 2), which is essentially an aggregation of all gestures/interactions performed, presented in a visual fashion, making it possible to zone in on user interface elements that users might find counterintuitive.

By learning what the usability, UX and performance issues are and making the right changes, it is possible to significantly improve KPIs, such as minimizing bounce rates, boosting retention, maximizing conversions, etc.

This analytics approach is mainly appropriate to product management, including feature optimization and release management, as well as UX designers.

Figure 2: Appsee UX Analytics’ heatmap (left) and user session recording (right), providing an in-depth analysis of user behavior. Source: Appsee website, available here

Behavioral analytics

To reveal the true digital actions of users, behavioral analytics is the tool of choice. The entire user journey is presented as a timeline (organised raw data) for each user’s behavior. In this way, marketing teams can see the preferences of users, see what they do not like and what they do not and optimize the product based on such data.

One can think of behavioral analytics as the true evolution of business intelligence — it is more than simply seeing where users click.

This type of analytics truly goes beyond standard analytics to uncover actionable insights such as which pages were viewed, where they were referred from, where they are geographically, where customers click and how many of them continue down the funnel to checkout.

Behavioral analytics relies on a plethora of user event data (see Figure 3), gathered from all customer actions (by user) across mobile, web and marketing campaigns, and tracks them as they access multiple devices, across multiple sessions. Behavioral analytics is built to track across diverse data sources and monitor digital behavior, presenting this behavior as a timeline of customer actions. This approach yields a meaningful understanding of what customers do.

Over time, behavioral analysis can provide answers to multilayered questions, such as:

• How long does it take a registered user to ask for help?
• Which items did users search for in comparison with the products they ended up adding to their cart?
• How do these results vary between users from different traffic sources?

Figure 3: Behavioral analytics dashboard provided by Mixpanel, which enables to have high and low-performing groups of users surfaced automatically that management can review, monitor and act on. Source: https://segment.com/integrations/mixpanel/

When one can analyse the sequence of events that users took over time, it is possible to build a much clearer picture of user behavior (the why and how). By breaking it down into smaller steps, one can identify the goals of each event and ask what actions drove the user to complete each action.

Behavioral analytics can be used for digital assets such as gaming apps and e-commerce sites to dig deep into the user experience, and by using those real-time insights, management can make business decisions immediately.

Moreover, management’s strategic use of actionable insights will allow for the mapping of their services and products. By mapping to their target audience’s core wants and desires, they are facilitating a key push behind success, both in customer satisfaction and maximized revenue.

For example, marketing management at a gaming company needs to make decisions based on open rates and increasing budget for campaigns. Instead of just using the traditional metrics that Google Analytics provides, they receive more detailed reports communicating in real time the rates and amounts of deposits or the installs deriving from the marketing campaign in Western Europe.

AI powered predictive analytics

In a world where only 2 per cent of website visitors make a purchase (REF3) solutions like Optimove and AgilOne make sense of the user’s entire journey from the very early stages, and when conditions are right, even before they have visited your website.

By combining artificial intelligence (AI) and machine learning (ML) and applying this to historic retention data, such analytics can produce models of future customers before they even make a purchase (see Figure 4). This allows marketers to target their audiences precisely, rather than rely on spraying and praying, which many do with programmatic advertising.

In the context of predictive analytics in marketing, the potential is tremendous.When applied to marketing, machines can learn how to classify consumers to help predict lifetime value. For management, access to such information can inspire actions to focus on particular types of acquisition campaigns that will bring more of a particular type of customer closer to the brand.

To achieve this customer ordering prediction feat could potentially minimize shipping, supply chain and inventory costs.

Logistics and supply chain optimization is neither cheap, nor is it simple. However, if companies want to put a dent into their costs while improving performance, this is the right opportunity to introduce some prediction magic.

Figure 4: Predictive micro-segmentation provided by Optimove, which applies mathematical and statistical models to transactional, behavioral and demographic data in order to predict future customer behavior and value. Source: Optimove website, available here

In the 2018 keynote of its Supply Chain Executive Conference in Phoenix, Gartner cited ‘automation, algorithms and the systematic use of real-time data as industry game changers’(REF4).

As one begins the task of turning AI-powered insights into dashboards, opportunities will be coming in at a faster and clearer pace, thereby empowering management to save the company time and money while streamlining decision-making processes for the organisation.

However, the power of AI does not slow down there. Mining for customers with similar attributes to the most loyal ones can tie in to user acquisition strategy:

• Lead scoring: As most salespeople would agree, no two leads are created equal. This is definitely the case when trying to prioritize leads. However, what predictive analytics can do is teach how to tie the actions of existing loyal customers to help grow future marketing efforts. By harnessing the power of predictive analytics, lead scoring shifts from becoming a checklist of criteria from the sales team to more of a clearer data-driven view of the target customer. The rules that are in play with predictive analytics feeding to a sound automation tool can score leads based on a combination of data such as demographic, psychological or behavioral. The process has proven to be quite effective in determining hot leads that should be passed to sales or entering a lead-nurturing campaign.
• Propensity to churn: Similarly, keeping guard over acquired and retained customers (the baseline) gets a whole lot simpler when applying the power of prediction analytics. Here is how- we all learn from our mistakes. It is widely accepted that past behavior is a decent indicator of future behavior. You are already armed with troves of historic customer data. The marketing analyst is instructed to identify the signs and behavior of customers who churned and supply report, which of course will come with a request for a higher budget for targeting those customer segments via a nurturing campaign focused on churn prevention.

The next phases of marketing analytics will only be pushing forward into overdrive as AI and ML take a bigger role in efficient data analysis and empower marketers to hyper-personalize their messages to customers. This includes providing them with unique journeys and personalizing content.

The consumer of today does not want to be treated as a part of some segment. Having supplied their data, they expect to be treated with a higher level of intimacy.

Marketing automation capabilities are already being transformed via AI and ML.

In the coming years, marketers will be harnessing the power of these game-changing technologies to help them with lead scoring. They will be offering even more personalized and dynamic content based on what phase of the user journey their customers are in, as well as setting up campaigns using complex triggers.

These initiatives will produce high converting experiences for those who get it right.

While most marketers are applying performance measurement in some respect, the bottom line is that the vast majority are not measuring all touch points and engagements. There is clearly a lack of visibility into customer behavior and this can lead to management being blindsided and misinformed, leading to poor management decisions.

According to the 2018 Gartner Marketing Analytics Survey, data analytics teams are spending ‘more time wrangling data than building insights’(REF5).

While this is not something for marketers to boast about, it presents marketing management with a huge opportunity to beat the competition by implementing some type of analytics now, before the competition gets on the analytics train.

DATA-DRIVEN MARKETING

According to Adobe, companies rely on data ‘as a foundation for all of their digital efforts’ (REF6). Data-driven marketers are not just utilizing historic and real-time data. As shown in Figure 5, they are merging such data with data in their CRM and other technologies in order to build a more complete insight generating ecosystem, as well as a more complete profile of their audience.

For example, CRM data are being used to strengthen analytics data by filling in the holes and creating a more complete buyer persona. Household income information obtained via CRM can influence a decision to offer higher-priced items to specific visitors.

While the use of CRM data remains the most common approach to augment data-driven marketing, the use of predictive analytics has grown by 36% year-on-year (REF7). Companies acknowledge the importance of predictive analytics: 69% feel that it is a ‘very important’ investment to make in the next three years. Meanwhile, budget-wise, half of organizations surveyed plan to increase spending on analytics, with some of this likely to be spent on real-time analytics.

Figure 5: Technologies that influence data-driven marketing — The chart reflects the various approaches used by digital marketers at enterprise organisations i2n order to positively complement their data-driven marketing. Image from MarketingCharts, based on data from Adobe; reproduced with permission. Source: MarketingCharts (2017) ‘How enterprise organizations are enhancing data-driven marketing’, available here

CONCLUSION

?Over the years, marketing analytics has taken businesses to varying levels of success. It is important to decide which flavor of analytics is most appropriate for management needs and diligently use the analytics platform to its full potential. AI-powered predictive analytics will definitely play a crucial role in marketing analytics and will take businesses to the next level.

---------------------------------------------------------------------------------------------------------------

Any questions? Facing a marketing challenge?

Feel free to reach out at [email protected]

---------------------------------------------------------------------------------------------------------------

References

1. Lewis, G. (2018) ‘How Hershey used data to increase retention rates and improve workforce planning’, available at: https://business.linkedin.com/talent-solutions/blog/talent-analytics/2018/howhershey-used-data-to-increase-retention-rates-andimprove-workforce-planning
2. The Appointments Bureau (2018) ‘Are you wasting money talking to the wrong people?’, available at: https://theappointmentsbureau.com/2018/09/14/wasting-money-talking-wrong-people/
3. ALC (n.d.) ‘You’re losing 98 per cent of your website visitors — how do you get them back?’, available at: https://www.alc.com/official-blog/98-percentwebsite-visitors-go-unnoticed-get-them-back/
4. Durbha, M. (2018) ‘Takeaways from the 2018 Gartner Supply Chain Executive Conference’, available at: https://blog.kinaxis.com/2018/05/takeaways-from-the-2018-gartner-supply-chainexecutive-conference/
5. Pemberton, C. (2018) ‘Key findings from Gartner Marketing Analytics Survey 2018’, available at: https://www.gartner.com/smarterwithgartner/key-findings-from-gartner-marketing-analyticssurvey-2018/
6. Adobe (2017) ‘Running on experience’, available at: https://landing.adobe.com/en/na/products/marketing-cloud/ctir-3108-running-on-experience/index.html
7. MarketingCharts (2017) ‘How enterprise organizations are enhancing data-driven marketing’, available at: https://www.marketingcharts.com/business-of-marketing-80902