Using real-time analytics, your conversations with customers can be more relevant and have a more personal touch. When you’re engaged in conversation, you may not be aware of it, but your brain is taking in and using massive amounts of information.
First, you recall previous conversations and everything you already know about the other person. You are also analyzing and building on what they are saying. Finally, depending on the nature and intensity of the relationship, your responses will be casual or formal in tone (or somewhere in between).
How much better would your conversations be if you could be more aware of the communication nuances as the conversation is occurring? And perhaps more importantly, what if you could do the same with your conversations with customers?
The answer for customer interactions is integrating real-time analytics into your business. By helping you to develop a comprehensive understanding of your customer, analytics will help you determine the next best offer, or at least the right message to send.
Using real-time analytics, your decision will be based on the context and history of interactions between the consumer and your organization – a complaint a few seconds earlier, a product ordered online or a basket abandoned. This dimension adds the immediacy necessary to any meaningful dialogue. SAS® Real-Time Decision Manager provides the technology and software to enable this. My colleague Andrea Sangalli has discussed this in his post dealing with the power of contextual marketing in real time.
When ‘Right Time’ is ‘Real Time’
The instantaneous nature of this information gathering and analysis is especially important in the connected world in which we operate. The marketer does not have the luxury of being able to take their time in responding to a prospect or customer. Incoming interactions, initiated by the consumer, override outgoing actions and are also much more precise and promising in terms of conversion. Now, it is the consumer who decides what he or she wants, when and by what channel.
For example, a customer logs into his bank account online to look at the interest rate on his saving account, then an associated FAQ page followed by the page that shows him the steps needed to move an account to another bank. This is a strong signal to the bank that it needs to take action. Additionally, he may call the bank’s call center to explore other options.
Predictive analytics will allow call center personnel to offer the bank’s best response to that customer based on his recent queries and previous history. It can also make sure that the incentive is appropriate for the size and importance of the account.
If analytics had been used without this real-time information, a next best offer would have been sent, but perhaps too late because the client would already have contacted the call center and the call handler, unaware of the online search for information, might have missed the opportunity to make the right offer. One customer lost – and how many more in future?
Rand Merchant Bank (RMB) ran an advert a few years ago, showing identical twins born 10 minutes apart. The advert shows how this small difference had a great bearing on their lives, with two very different personalities developing, and, of course, both eventually finding contrasting but equally fulfilling careers at the bank.
I was reminded of this recently when discussing the differences in behaviour within very similar customer segments. No matter how granular and analytically-advanced your segments, there are minor differences between the individuals in each one that cause them to behave differently from their peer group. In the words of the advert, they are “identical in every way, except the way they think."
Data-driven marketers are trying to overcome this problem with a segment-of-one approach. Listening to, understanding and then acting on the unique nuances in behaviour of each individual means truly personalised experiences.
Will we be seeing truly personalised experiences in the near future?
Imagine an online fashion retailer. Two identical male shoppers, in the young professionals segment, interested in smart suits and accessories, both shopping online, both presented with the same offers for the latest Italian designs . . . but one day, one of these shoppers clicks on the kids’ clothing section, not the men’s.
Is this a mis-click? Or does this young man have a new niece or nephew, or maybe a newborn son we didn’t know about?
And more importantly, what does this mean for the way that we handle his visits and make offers to him? Do we move him into the new parents segment? How do we know if this is a permanent change in behaviour? Can we triangulate this behaviour with any other that we’ve seen, to provide clues?
The technology is available to do this. We have big data processing power and the analytical capabilities to sift through this data to uncover relevant patterns of behaviour. We can even do it in real time, or close to it.
But the complexities in data collection and integration slow these efforts down. Companies are overwhelmed by the volume of the data, and struggle to identify common threads across multiple sources. Organisational silos, channel isolation and segment-based thinking all hamper company-wide efforts to develop the elusive 360-degree view of individual customers that would allow real-time analysis of their behaviour.
How can organisations realign around a segment of one?
I suggest that customer-centricity is developing a new meaning. It is now understood that realigning the organisation’s data, people, processes and technology around its customers is the only way to achieve truly personalised experiences. It is also understood that these experiences will be the cornerstone to winning and retaining customers.
But there is another problem. Even if an organisation could collect all its data and analytics in one place and build an intelligent view of each customer’s unique past behaviour, it would struggle to react quickly enough to nuances in real-world behaviour. And this is what is required for true personalisation.
The most vital component for marketers and data analysts is a centrally-managed data, analytics and real-time decision-making engine at the heart of all marketing efforts. This centralised engine should act as the channel-agnostic and context-sensitive brain. It would be working in the background during all interactions across all channels and make real-time decisions for these channels about what messages to provide to each customer.
Many organisations make the mistake of building personalisation logic, but limiting it to a particular channel, usually the website or mobile channels. If these channels operate in isolation and don’t listen to (or feed) the centralised brain, insights and decisions made on this channel do not inform, nor are informed by, any other channel, whether it’s the call centre, in-branch staff or batch email marketer.
The work of the centralized brain
Let’s go back to our example. Our young male customer’s recent change in behaviour cannot be handled in isolation. As soon as this behaviour occurs, the central engine should move into action.
It is constantly listening for new contextual information, such as website or mobile app clickstream data. When it obtains new data, it runs it through a real-time process to decide if this new information should change our predetermined action for this person. This process considers all available data, such as:
Engagements with the brand in the past minutes or hours (since the last batch analytical processes ran).
Insights on social media using text analytics.
Previous browsing history to check whether this is an isolated incident.
Purchase history to see if he does this at the same time every year.
At the end, the brain will make a decision about whether to override or append the predetermined scores or segments. It will determine the best action to take for that individual in that moment (our segment of one), and this action will immediately be available to all other channels, brands and data sources.
Is this really worth all the effort, time and expense? Well, SAS’ customers think it is. One mobile operator is able to detect real-time context in airtime balance thresholds. Their problem was that they could not send personalised offers until a few hours after the threshold was reached. And by then, the offers were often no longer relevant. The company was hovering at a 5 percent response rate to its offers no matter what it tried.
But when the company introduced real-time centralised decision-making with personalisation, response rates rose to 24 percent, generating tens of millions in incremental revenue per year. This was far beyond expectations and will only improve as the company’s capabilities mature.
Time to change
Is your organisation treating your customers like they are all twins? Improving your segmentation abilities is an evolutionary journey, and I urge you to start immediately with what you have. The white paper by Suneel Grover, Analytics in Real-Time Online Marketing, discusses how your organisation can take the first steps to detect, analyse and respond to the rich data that your customers are already giving you on digital channels.
This SAS eBook is another great primer on the concept of contextual marketing.
Consider the last email or digital ad you received from a favorite retailer. It may have included an offer to save 20 percent on your next online purchase, or an invitation to shop in store during an exclusive sale. You don't think too much about this brand’s customer relationship management (CRM) or marketing capabilities, because you don’t have to.
Why? Because the most sophisticated brands employ tools that can tailor an email or a social media post to their buyer’s sweet spot. Powered by data and analytics, these CRM tools do the heavy lifting for marketers to engage their customers in more personalized, authentic ways.
CRM Watchlist 2016
Often recognized as a forerunner in CRM software, SAS Customer Intelligence has added a new accolade to its trophy case as a winner on the 2016 CRM Watchlist. The annual list – curated by leading CRM industry analyst, Paul Greenberg – includes the dominant companies to watch in the CRM market. As Greenberg notes in his announcement blog post on ZDNet, the competition was especially stiff this year, with 131 vendors vying for the winners spot. With each submission, Greenberg reads and scores the company (which is weighted), which is then followed up with extensive research analyzing the vendor in the markets it addresses.
One important distinction of the Watchlist is the winner's impact within the CRM space. Greenberg cites that “the impact has to be obvious, both in the prior year and in the anticipated next two or three years.” And “that there is no doubt at all that your company is making a major impression on a market and actually changing or strengthening that market by its presence.”
The impact is not only from the strength of our SAS customer intelligence offerings, but as a whole company. Greenberg states, “To have an impact, the company has to be pretty much a complete company who has been doing this long enough to have established a rhythm that leads to impact. The company has to be well rounded -- it has financial stability, solid management, excellent products and services, culture, and a strong partner ecosystem to help sustain its efforts.”
The SAS customer intelligence team is honored to earn a spot on the winners list for 2016, demonstrating SAS's commitment to helping brands deliver customer experiences that matter.
Today’s retailers have access to vast stores of data that allow them to create the personalised retail experience that customers have come to expect. Used in the right way, analytics can be the key to bringing customers in through the door, building a better online experience, or simply helping weather slow periods by enabling faster, more efficient and nimble supply chain changes.
Interestingly, customer perceptions towards a brand can significantly change following Black Friday discounting and January sales. Pricing strategy can be confusing at the best of times. Given the option of £50 off on a high-value product compared to a 70 percent discount on a low value item: which is the better offer? Which deal do consumers trust, and why? More importantly, how does it impact the brand perception?
One thing’s for sure: brand loyalty and consumer trust are no longer guaranteed in today’s fast-changing retail environment. Retailers must find alternative ways to engage with consumers.
The dark art of forecasting
For years, retailers have collected customer data through loyalty cards, email marketing promotions, point-of-sales tills, as well as online browsing behaviour and purchase orders. What has changed over the years is the technology that allows retailers to analyse and understand data. The advent of technologies like Hadoop has enabled the development of advanced analytics solutions to produce more insightful and timely answers for retailers, which is of huge value during peak trading seasons.
Another interesting point here is that often the data collected spans the last three years of trading or more. This is traditionally used by store managers to forecast pricing and demand. Yet we know that outdated historical data, coupled with unpredictable external factors like weather conditions, regularly produce inaccurate forecasts.
As a result, retailers are still making predictions largely based on gut feel rather than objective insights. More often than not, they are also spontaneously reacting to competitors’ price cuts without carefully calculating their profits and loss potential.
Brands switch off
Another common retail pitfall is the effect of over marketing with a general blanket message. Unaware of the consequence of blasting out daily and sometimes twice-daily promotional emails in the hope of catching shoppers’ bargain hunting instincts, retailers are actually turning customers away. Consumers will easily unsubscribe from brands they previously loved, especially if they feel they are bombarded with irrelevant and unwanted product recommendations that only clog up their inboxes.
True personalisation demands an intimate understanding of the customer, and willingness from the individual to participate.
One of our retail customers has been using data analytics to deliver personalised product displays. These are highly targeted to improve conversion and avoid endless page scrolling. Instead of feeling like every other shopper each time a customer enters the site, they can now enjoy a genuinely personal engagement with the retailer. Behaviour is predicted and product recommendations are based on browsing, and a wealth of customer and transaction data.
With a more personalised customer experience, shoppers will be able to purchase products at the best price, while receiving the best in-store, online and post-sales service.
As retailers have better insights into customer behaviour and spending patterns, they will be able to personalise the experience and product recommendation. In turn, shoppers will feel they are receiving better deals and being looked after by the retailer. Some customers may also find themselves receiving extra benefits or upgrades through loyalty schemes, enhancing the overall experience with the brand.
SAS has worked for many years with dunnhumby, the power behind Tesco’s Clubcard. Today there is a need to market to customers across multiple channels. SAS has also worked with Callcredit Information Group, specialists in marketing services, analytics and data, to deliver a market leading omnichannel marketing and analytics service. This partnership has delivered a solution to ASDA. It provides an easy-to-use interface, with repeatable processes to enable organisations to deliver more campaigns – from simple to complex – in a shorter space of time.
Sail through the online retail revolution
On Black Friday alone, shoppers spent £1.1 billion with UK online retailers, setting a new internet retail record which underlines the online shopping revolution. The true impact of this year’s trading results are yet to be seen but one thing is certain – retailers can no longer rely on historical data alone to forecast pricing models or stock performance.
The cost for retailers can be drastic if they cannot accurately analyse the net profitability impact ahead of the peak trading period.
Big data enables retailers to more accurately forecast actual consumer demand, based on the very latest fashions and trends, as well as the timing and the likely location for that demand. For example, big data can be used as follows:
Analyse the data and visualise the demand forecast on a specific product in a specific store on a specific time and day of the week. This enables the retailer to ensure they get specific stock to that location.
Predict what else the customers might like to buy in the same transaction and personalise the product recommendations accordingly.
Predict the behaviour of customers by channel. For example, how customers might move between online browsing and in store purchasing, or their preference for home deliveries or Click & Collect.
Use demand insight to negotiate more effectively what the retailer buys from suppliers and when. Also, strategically plan for a more coherent and cheaper supply chain and transportation journey, allowing for external factors like extreme weather conditions.
Understand which segment of the customer demographics are most valuable to the business and devise a more effective way to nurture their spending and relationship with the brand.
Delight your customers by presenting them with product and brand recommendations which you already know they will want and like.
In principle, the knowledge of who wants what and when is an art form in retailing that is rooted in the golden era of advertising and knowing your customer. Retailers that have enhanced their skills in this area will continue to grow and prosper, which is why demand for data analytics technology will continue to grow over the coming years.
Martech is moving fast - pulling in other segments of technology as it pushes forward into 2016. We are well on the way to 2020 - where WalkerInfo predicts that customer experience will be the primary differentiator for consumers over the traditional four P's of marketing (product, place, price and promotion). In addition, a Salesforce survey of 5000 B2B and B2C marketers shows that keeping up with trends is one of the biggest problems marketers face today. While technology and consumer behavior are involving in unpredictable ways, so are the ways marketers are reacting to these new developments. Here are a few things that I see coming in marketing’s future, and I think they will be here before we know it.
Future trend 1: Nonhuman marketing
Today, businesses have more data than ever before, including network and device data, collecting from proliferating channels. Our televisions, phones, Fitbits and Apple Watches are becoming connected, rapidly changing consumer behavior. For example, the multiscreen trend in which people watch television while also using other digital devices presents a great opportunity for marketers. Combining this data with traditional demographic and geographic consumer data can greatly improve marketing efforts.
All of this connected technology will become platforms for my life, creating a myriad of consumer data and better experiences. With this trend, marketing will need to direct its attention to all devices, and not just to individuals. SAS is combining streaming analytics off all of these device types with traditional data like demographic, social, and transactional data to help brands inform marketing efforts and send marketing messages and offers to channel touch points.
Future trend 2: Consolidation
Businesses strive to achieve a single customer view because they understand the value to their bottom line, including making marketing more effective and gaining better insight into customers to create brand value and greater loyalty. Brands are starting to understand that customer experience makes the sale, not product or price. By removing intermediaries, supply chains are shorter and technology becomes more user-enabling.
Today, some businesses serve the sole purpose of aggregating and sourcing products and services for consumers. At amazon.com, I can buy anything from a TV to a pair of socks in a single click – without ever interacting with Samsung or Woolrich as brands. So who do these brands market to? Not me the consumer – but Amazon as the aggregator – and the message to that aggregator will obviously be much different than the marketing messages of today. SAS is working in a B2B context with many brands to understand how and who to market to in order to break through the noise and deliver offers to businesses and intermediaries as we see business models change as we move into 2016.
Future trend 3: Applied analytics and inherent intelligence
I talk to people about marketing software technology every day. Never once have I heard someone say – “I really love my marketing technology software – I wish I could spend more time in it, you know,just clicking around and hanging out.”
Instead, marketing analysts want to quickly enter the software, perform an action, complete a task and exit the application so that they can get on to more strategic parts of their job. Marketing technology vendors are answering marketers’ desire for faster and more intuitive software.
The ability to embed and apply analytical techniques – things like automatically derived segmentation, applied marketing activity optimization, embedded forecasting, and other techniques – all in an effort to infuse machine learning style techniques and behaviors into marketing interfaces. SAS is working to infuse these cognitive analytical behavior techniques into marketing technology software. After all, every marketer (including myself) wants to simplify and optimize workloads! If a machine can tell a marketer the best time to execute a campaign, to whom, which content to use, and when to send it out – why wouldn’t a marketer oblige? I know I would.
I begin this blog post with one goal in mind. I want to raise awareness on the subject of customer and marketing analytics, and why this field is exploding in interest and popularity. Let's begin with a primer for the uninitiated, and lay down some definitions:
Customer Analytics: The processes, technologies, and enablement that give brands the customer insight necessary to provide offers that are anticipated, relevant and timely.
Marketing Analytics: The processes and technologies that enable brands to assessthe success of their marketing initiatives by evaluating performance using important business metrics, such as ROI, channel attribution, and overall marketing effectiveness.
If you aren't a fan of textbook definitions, here is a creative alternative:
Still not on board? Here's my perspective on the subject:
Customers are more empowered and connected than ever before, with access to information anywhere, any time – where to shop, what to buy and how much to pay. Brands realize it is increasingly important to predict how customers will behave to respond accordingly. Simply put, the deeper your understanding of customer buying habits and lifestyle preferences, the more accurate your predictions of future buying behaviors will be.
Marketers need to be enabled to benefit from approachable and actionable advanced analytics to make more powerful decisions within today’s complex and interconnected business environments. In my mind, the big picture boils down to one, two or three core enablers, based on your organization's goals and preferences:
Marketing analysts tasked with making sense of customer data, big or small, have to migrate through a complex maze of myths and realities about technology platforms, advanced analytics solutions and, most importantly, the magnitude of customer analytics efforts. On the surface, it appears that customer analytics is a well-entrenched discipline in many organizations, but under the hood, old problems persist around data integration and data quality while new ones emerge around the real-time application of insights and the ability to rein in digital data for customer-based analysis.When I speak with clients, there are two key themes that I continually hear:
Data is a big challenge. As customer interactions with brands increase and diversify, brands need to integrate data effectively in order to provide the contextual and real-time insights their customers are growing to expect. Haven't you grown tired of saying we spend 80 percent of our time on data management related tasks, and 20 percent on analysis?
Analytic talent is hard to find. Brands struggle to find individuals with the right analytic skills to meet the challenges they are facing today. Without the talent to unlock actionable insights, modern customer analytics cannot meet its potential. (Given my public affiliation with The George Washington University's M.S. in Business Analytics program, I'd recommend checking it out if you are hunting for quality talent.)
To me, these themes point to a workflow entitled the marketing analytics lifecycle:
With the growing importance of customer analytics in organizations, the ability to extract insight and embed it back into organizational processes is at the forefront of business transformation. However, this requires considerations for where relevant data resides, the ability to reshape it for downstream analytic tasks (predictive modeling vs. reporting), and how to take action on the derived insights. Furthermore, there are the roles of different people within the organization that need to be considered:
Supporting IT Team
Customer analysis touches all of these roles, and to enable this audience comprehensively, all aspects of the marketing analytics lifecycle must be supported. To directly address this, I want to to highlight what SAS is doing to help our clients meet these challenges.
Marketing Analytics Lifecycle Stage #1: Integrate and Prepare Data
Customer analytics is highly dependent on the quality of the ingredients we feed into analysis. Now, the digital marketing industry has been taken by storm by the emergence of Digital DMPs, like Oracle BlueKai, Neustar, and Krux, who aim to provide marketers support in programmatic ad buying and selling. Marketers and publishers are learning that harnessing their first-party data; developing single and consistent identities for their consumers across devices and systems, like email and site optimization; and gaining access to second-party data are mission critical. However, the subject of data mining and predictive analytics has largely been ignored by the Digital DMP space. Brands who want to exploit the benefits of advanced analytics have additional considerations to support their data management challenges. The following video highlights how SAS helps manage and prepare data of all sizes, from 1st party customer data to clickstream and IoT, specifically for analytics:
Some of you might be questioning the value of this, so let me offer a different perspective. Over the past few years, I have developed a personal frustration of attending various marketing conferences and repeatedly observing high-level presentations about the potential of analytics. Even more challenging has been the recent trend of companies presenting magical (i.e., "easy-button") black-box marketing cloud solutions that address every imaginable analytical problem; in my opinion, high-quality advanced analytics has not reached a point of commoditization, and remains a point of competitive differentiation. Do not be mislead by sleight-of-hand magic!
What types of marketing challenges are you attempting to solve with customer analytics? Srividya Sridharan and Brandon Purcell are two leading researchers in the space of customer insights, and recently released a report entitled How Analytics Drives Customer Life-Cycle Management recommending the deployment of various analytical techniques across the customer life cycle to grow existing customer relationships and provide insight into future behavior. Highly recommended reading! Let's review some of the most common problems (or opportunities) we view at SAS with our clients.
Within each of the categories, a myriad of analytic techniques can be executed to assist and improve your brand's abilities to address them. The following video is a demonstration of how I used SAS Visual Statistics and Logistic Regression analysis to understand drivers by marketing channel of business conversions on a website or mobile app. The benefit of understanding these data-driven drivers is to influence downstream marketing personalization and acquisition campaigns. In addition, capabilities related to group-by modeling, deployment scoring and model comparison with other algorithmic approaches are highlighted.
Big digital data, scalable predictive analytics, visualization, approachability, and actionability. Stay thirsty my friends, because it is our clients who are expressing their needs, and SAS is stepping up to meet their challenges!
If you would like to learn more on how we address other marketing and customer analytic problems, please click on any of the following topics:
With that said, we have one final stage of the lifecycle to review.
Marketing Analytics Lifecycle Stage #3: Explain Results and Share Insights
An individual's ability to communicate clearly, succinctly and in the appropriate vernacular when presenting analytical recommendations to a marketing organization is extremely important when focused on driving change with data-driven methods. I recently wrote a blog post on this topic entitled Translating Predictive Marketing Analytics, and if you're tired of reading, here's another video - this time focused on explaining the results of analytical exercises in easy-to-consume business language.
As I close this blog post, I want to leave you with a few thoughts. For your brand's customers, technology is transparent, user-enabling, and disintermediating. The journey they embark with you on is fractured and takes place across channels, devices, and points in time. The question becomes – are you prepared for moments of truth as they occur across these channels over time? Customer analytics represents the opportunity to optimize every consumer experience, and revisiting a point I made earlier, the deeper your understanding of customer buying habits and lifestyle preferences, the more accurate your predictions of future buying behaviorswill be.
If you enjoyed this article, be sure to check out my other work here. Lastly, if you would like to connect on social media, link with me on Twitter or LinkedIn.
Marketers have used segmentation as a technique to target customers for communications, products, and services since the introduction of customer relationship management (i.e., CRM) and database marketing. Within the context of segmentation, there are a variety of applications, ranging from consumer demographics, geography, behavior, psychographics, events and cultural backgrounds. Over time, segmentation has proven its value, and brands continue to use this strategy across every stage of the customer journey:
Let's provide a proper definition for this marketing technique. As my SAS peer and friend Randy Collica stated in his influential book on this subject:
"Segmentation is in essence the process by which items or subjects are categorized or classified into groups that share similar characteristics. These techniques can be beneficial in classifying customer groups. Typical marketing activities seek to improve their relationships with prospective and current customers. The better you know about your customer's needs, desires, and their purchasing behaviors, the better you can construct marketing programs designed to fit their needs, desires, and behaviors."
"In an era of big data, hyperconnected digital customers and hyper-personalization, segmentation is the cornerstone of customer insight and understanding across the modern digital business. The question is: Is your segmentation approach antiquated or advanced?"
This provides a nice transition to review the types of segmentation methods I observe with clients. It ultimately boils down to two categories:
Business rules for segmentation (i.e., non-quantitative)
Analytical segmentation (i.e., quantitative)
Let's dive deeper into each of these...
Business Rules For Segmentation
This technique centers on a qualitative, or non-quantitative, approach leveraging various customer attributes conceptualized through conversations with business stakeholders and customer focus groups to gather pointed data. This information represents consumer experiential behavior, and analysts will assign subjective segments for targeted campaign treatments. Although directionally useful, in this day and age of data-driven marketing, it is my opinion that this approach will have suboptimal results.
Within this category, there are two approaches marketing analysts can select from:
Supervised (i.e., classification)
Unsupervised (i.e., clustering)
Supervised segmentation is typically referred to as a family of pattern analysis approaches. Supporters of this method stress that the actionable deliverable from the analysis classifies homogeneous segments that can be profiled, and informs targeting strategies across the customer lifecycle. The use of the term supervised refers to specific data mining (or data science) techniques, such as decision trees, random forests, gradient boosting or neural networks. One key difference in supervised approaches is that the analysis requires a dependent (or target) variable, whereas no dependent variable is designated in unsupervised models. The dependent variable is usually a 1-0 (or yes/no) flag-type variable that matches the objective of the segmentation. Examples of this include:
Product purchase to identify segments with higher probabilities to convert on what you offer.
Upsell/cross-sell to identify segments who are likely to deepen their relationship with your brand.
Retention to identify segments most likely to unsubscribe, attrite, or defect.
Click behavior to identify segments of anonymous web traffic likely to click on your served display media.
After applying these techniques, analysts can deliver a visual representation of the segments to help explain the results to nontechnical stakeholders. Here is a video demonstration example of SAS Visual Analyticswithin the context of supervised segmentation being applied to a brand's digital traffic through the use of analytical decision trees:
Critics of this approach argue that the resulting model is actually a predictive model rather than a segmentation model because of the probability prediction output. The distinction lies in the use of the model. Segmentation is classifying customer bases into distinct groups based on multidimensional data and is used to suggest an actionable roadmap to design relevant marketing, product and customer service strategies to drive desired business outcomes. As long as we stay focused on this premise, there is nothing to debate.
On the other hand, unsupervised approaches, such as clustering, association/apriori, principal components or factor analysis point to a subset of multivariate segmentation techniques that group consumers based on similar characteristics. The goal is to explore the data to find intrinsic structures. K-means cluster analysis is the most popular technique I view with clients for interdependent segmentation, in which all applicable data attributes are simultaneously considered, and there is no splitting of dependent (or target) and independent (or predictor) variables. Here is a video demonstration example of SAS Visual Statistics for unsupervised segmentation being applied to a brand's digital traffic (including inferred attributes sourced from a digital data management platform) through the use of K-means clustering:
Keep in mind that unsupervised applications are not provided training examples (i.e., there isn't a 1-0 or yes/no flag type variable to bias the formation of the segments). Subsequently, it is fair to make the interpretation that the results of a K-means clustering analysis is more data driven, hence more natural and better suited to the underlying structure of the data. This advantage is also its major drawback: it can be difficult to judge the quality of clustering results in a conclusive way without running live campaigns.
Naturally, the question is which technique is better to use in practice – supervised or unsupervised approaches for segmentation? In my opinion, the answer is both (assuming you have access to data that can be used as the dependent or target variable). When you think about it, I can use an unsupervised technique to find natural segments in my marketable universe, and then use a supervised technique (or more than one via champion-challenger applications) to build unique models on how to treat each cluster segment based on goals defined by internal business stakeholders.
Now, let me pose a question I have been receiving more frequently from clients over the past couple of years.
"Our desired segmentation strategies are outpacing our ability to build supporting analytic models. How can we overcome this?"
Does this sound familiar? For many of my clients, this is a painful reality limiting their potential. That's why I'm personally excited about new SAS technology to address this challenge. SAS Factory Miner allows marketers to dream bigger when it comes to analytical segmentation. It fosters an interactive, approachable environment to support working relationships between strategic visionaries and analysts/data scientists. The benefit for the marketer campaign manager is the ability to expand your segmentation strategies from 5 or 10 segments to 100's or 1000's, while remaining actionable within the demands of today's modern marketing ecosystem. The advantage for the supporting analyst team is the ability to be more efficient, and exploit modern analytical methods and processing power, without the need for incremental resources.
Here is a video demonstration example ofSAS Factory Miner for supersizing your data-driven segmentation capabilities:
I'll end this posting by revisiting a question we shared in the beginning:
Is your segmentation approach antiquated or advanced?
Dream bigger my friends. The possibilities are inspiring!
If you enjoyed this article, be sure to check out my other work here. Lastly, if you would like to connect on social media, link with me on Twitter or LinkedIn.
I get the most interesting insights from questions my kids ask me about my work. Why? Because they know very little about big data or analytics, and the questions they ask are sometimes about things that I’ve taken for granted. Every time that happens, it reminds me that the questions you ask are just as important as the answers you get, and that different perspectives are often needed to see the whole picture of a situation.
Photo: Barry Butler, Chicago, IL, taken in Botswana.
That last point is colorfully laid out in the old Indian story of the Blind Men and an Elephant, which is described in Wikipedia as:
A group of blind men (or men in the dark) each touch an elephant to learn what it is like. Each one feels a different part, but only one part, such as the side or the tusk. They then compare notes and learn that they are in complete disagreement.
The issue, of course, is that none of them have all the information to describe the entire beast.
The elephant story and my kids’ questions also remind me of a lesson I learned years ago using SAS in a college econometrics class – don’t ignore the outliers in your model for two good reasons:
Outliers highlight potential flaws in your model – or aspects of your data that your model does not do a good job of explaining.
Outliers may be the leading indicators of a new trend that will completely upend your original model.
So the opportunity before us this coming year relates to elephants – how we’ll recognize them and what we’ll do with them once we see them. And one thing is for sure – we all have elephants. The key is to find them and to recognize the opportunities they represent. Read More »
Although the title of this blog posting has all the ingredients to attract the eyes of an analyst, the content is targeted for all personalities of a digital marketing organization. Before we jump into the marketing analytic use case regarding forecasting, scenario analysis, and goal-seeking for digital analytics, let's spend some time on the magic of stories. As Tom Davenport stated in his fantastic article titled, Telling a Story with Data:
"The essence of analytical communication is describing the problem and the story behind it, the model, the data employed, and the relationships among the variables in the analysis. When the relationships among variables are identified, the meaning of the relationships should be interpreted, stated, and presented relevant to the problem. The clearer the results presentation, the more likely that the quantitative analysis will lead to decisions and actions—which are, after all, usually the point of doing the analysis in the first place."
While creative visionaries and data scientists are both tremendous organizational assets within a team, it is the alliance between these two segments that will push marketing forward. Although aspirational, this is a difficult challenge to overcome. Let me begin by sharing a bit of my story - one that began with a four year career start in graphic design and creative marketing communications, and then making a leap to the quantitative side of marketing. I've seen and listened to how DIFFERENT these two segments of the marketing world are, and now as a preacher for the potential of marketing analytics, one's ability to make analysis interpretable and approachable is critical.
"The true value of data emerges when marketers are able to use it to tell a meaningful story. Enter the data storyteller, or marketing measurement analyst. This is the person who can push the tools, translate insights across the business, and motivate stakeholders to participate."
This quote nails the crux of the issue - if we don't take ACTION on the insights of analytics, it was nothing but a school project. Influencing decision-makers within an organization isn't easy, and if they do not understand the analysis, nothing will ever change. There are people who are good at creative marketing strategy, and there are people who are good at marketing analytics. However, there aren't many people who can toggle between the two, and serve as the translator who inspires both sides. Read More »
Broadly speaking, the holy grail of digital media measurement is to analyze the impact and business value of all company-generated marketing interactions across the complex customer journey. In this blog post, my goal is to take a transparent approach in discussing how data-driven marketers can progress past rules-based attribution methods, and make the business case for leveraging algorithmic applications.
Let's begin with a video example that pokes humor at the common problems related to multi-channel marketing attribution. The business challenge is that everybody believes they should have more marketing budget because their tactics are supposedly responsible for driving sales revenue. The video suggests that challenges arise rapidly when supporting analysis to justify these claims isn't sound. While the video is fictional, the problems are very real. With that said, there are three main drivers to getting digital attribution analysis right:
Allocating credit across marketing channels more accurately
Providing invaluable insights to channel interactions
Empowering marketers to spend more wisely in future media activity
Have you ever given thought to the many ways that a customer can find your brand's digital properties? Organic results on a search engine, display media campaigns, social media links, re-targeting on external sites, and the list goes on in today's fragmented digital ecosystem. One thing is for certain - consumer digital journeys are far from linear. They can occur across multiple platforms, devices and sessions, and organizations are challenged with gaining an accurate understanding of how:
External referral clicks (or hits) are mapped to channels and visits
Visits are mapped to anonymous visitors
Anonymous multi-channel visitor journeys are mapped to identifiable individuals across different browsers and devices
With careful consideration towards the areas of data management, data integration, and data quality, analyzing customer-centric (or visitor-centric) channel activity on their journeys to making a purchase with your brand can have immense benefits. Ultimately, marketers desire to apply a percentage value that can be attached to each channel's contribution to the purchasing event (or revenue). This is critical, as it allows the organization to determine how important each channel was in the customer journey, and subsequently, influence how future media spend should be allocated.
"There are few things more complicated in analytics
(all analytics, big data and huge data!)
than multi-channel attribution modeling."
The question is...why is it challenging? Avinash's blog post was written in the summer of 2013, and I strongly believe 2.5 years later we are living in a game-changing moment within digital analytics. Marketers are being enabled by technology companies with approachable and self-service analytic capabilities, and this trend directly impacts our ability to improve our approaches to problems like attribution analysis. However, rules-based methods of attribution channel weighting continue to be far more popular in the industry to date, which contradicts the recent analytic approachability trend. Before we dive into algorithmic attribution, let's review the family of approaches commonly applied in rules-based attribution:
Welcome to Customer Intelligence, a blog for anyone who is looking for ways to improve the business of marketing and communicating with customers.
We strive to prompt new thinking in the way you tackle customer-related business issues. And we hope to inspire the use of analytics for everything from multi-level marketing to social media campaigns. Follow us here and on Twitter at @SAS_CI, or check out the Twitter hashtag #sasci.