Predictive analytics is probably the hottest thing in marketing analytics right now. Predictive analytics go beyond describing consumer behavior to predicting how consumers will behave in the future based on data.
If you’ve never heard of it, don’t feel bad–marketing lags behind many other business disciplines in using data to drive decision-making. In fact, marketing still struggles with descriptive statistics, which merely describe behaviors, such as visits, ad responses, etc.
Marketers shouldn’t feel bad about their reliance on squishy analytics, such as increased awareness. In the past, marketers just didn’t have access to data, which was locked in the minds and hearts of consumers. Gaining access to this data was very expensive and time-consuming, forcing marketing managers to use less accurate and nuanced data such as customer demographics and point of sale data, according to this infographic by Absolutdata and alteryx. But, with digital marketing, marketing managers now have access to a plethora of valuable data and many
But, with digital marketing, marketing managers now have access to a plethora of valuable data and many struggle with how to use it although that access might be restricted to data analysts and they may lack the skills necessary to draw insights from the data.
So, today, I’d like to share some strategies for using predictive analytics, which, are exactly what they sound like. They predict behavior, rather than simply reporting on it.
What is predictive analytics?
We talk about 3 types of analytics:
Today, we’ll focus on predictive analytics, but if you’re interested in learning about the other 2 types of analytics and how they can help improve ROI, check out the links above.
Put simply, predictive analytics, as their name implies, try to predict the future, while descriptive analytics describes the past and prescriptive analytics help plan the best course of action (for instance, what are the best routes for buses in a major city).
Companies struggling to understand or implement descriptive analytics are now challenged to adopt this new form of analytics. A big part of the problem is data and the skills to analyze it is restricted to a limited number of data scientists who, often, lack the conceptual background to know what they’re looking for when it comes to predicting consumer behavior.
Benefits of predictive analytics
According to SPSS:
Organizations that incorporate predictive analytics into their daily operations in this way improve their business processes, enhancing decision making and gaining the ability to direct, optimize, and automate decisions, on demand, to meet defined business goals.
And, based on survey data, Forbes says:
A vast majority of executives who have been overseeing predictive marketing efforts for at least two years (86%) report increased return on investment (ROI) as a result of their predictive marketing.
Only 5% say they’ve not experienced an improvement in ROI or a decline in ROI from their efforts at predictive analytics.
Uses of predictive analytics in marketing
Arguably, predictive analytics support most enterprise activities, but I’m focused on marketing. Here are some uses for predictive analytics in marketing:
Segmentation involves separating a market into subgroups with similar attitudes, demographics, geographics, or behaviors. After segmenting your market, you position your product to appeal to the wants and needs of the chosen segment(s)–your target market.
Data aids in crafting your target segment(s) and determining the most effective positioning for each. Predictive analytics also helps to identify the most profitable segments based on historical consumer behavior within each segment.
Marketing managers use this data to allocate resources to reach the most profitable segments.
According to HBR, the biggest use of predictive analytics is in developing demand models that forecast sales and revenue–the starting point for budgeting.
3. Demand pricing
Demand pricing, often called yield management, involves pricing products based on differences in elasticity of demand between consumer groups. For instance, business travelers are willing to pay more for a seat on an airplane than casual travelers, so you can charge them more and reduce the price to casual travelers to make your flights more competitive and still meet ROI goals.
Using predictive analytics, firms conduct a series of experiments to determine factors affecting the impact of price on demand. Using these predictive models, firms develop optimal pricing strategies that maximize ROI.
4. Improve customer satisfaction
Customer satisfaction greatly impacts retention and loyalty. It also improves other positive consumer behaviors, such as recommending the brand to others. Any improvement in customer satisfaction impacts ROI, potentially significant. That’s the entire basis for the rise of relationship marketing, where marketing shifts its focus to pleasing existing customers than attracting new ones. Data suggest that it’s 5X less expensive to keep an existing customer than replace that customer.
Using data from customer service calls, mentions on social media, etc offer insights into factors leading to poor customer satisfaction. Using predictive tools such as conjoint analysis allows firms to discover which product improvements generate the greatest improvement in customer satisfaction.
Challenges to implementing predictive analytics
Looking at the infographic, we see that 43% of those surveyed feel that gaining access to the data is the biggest problem affecting their ability to use the data. Much of the data is housed in numerous functional silos, units of the organization, or across country borders making it difficult to get a clear picture of how customers respond to marketing efforts. Another 39% of respondents report difficulties integrating data across various places where it is stored. Creating actionable insights from the data is reported with only 37% of respondents. My guess is that this number.
Creating actionable insights from the data is reported with only 37% of respondents. My guess is that this number is even larger, in reality, and a function of the old adage “you don’t know what you don’t know”. Using data mining techniques mindlessly doesn’t generate insights and may obfuscate true relationships hiding in the data. Instead, knowledge of consumer behavior should guide analysis by identifying variables likely to impact important behaviors.
And, this is the real difficulty. Marketers aren’t often trained in data analysis techniques and the colleges and universities firms count on to teach these skills shrink from the job because they fear losing students or getting poor evaluations from students who hate analytics and math. To mitigate this problem, firms sometimes hire engineers who possess sufficient analytical skills, not realizing the true impact of their lack of knowledge about marketing concepts.
An obvious solution is to improve the analytical skills of marketers either by paying to train marketers or by hiring marketers with analytics training (which is easier said than done according to the infographic above showing 57% of firms have trouble hiring employees with analytical skills).
Alternatively, firms can buy software solutions that don’t require high-level analytical skills, such as those produced by the authors of the infographic.
That still leaves the issue of data spread all over the organization. Solving this problem requires firms to develop comprehensive data collection and storage plans that ease integration across different units. Combining data from different databases requires a key to link the data together as well as skills such as SQL to manage the integration. Again, software exists that allows firms to integrate with little technical knowledge.
Predictive analytics in digital
Digital marketing (or internet marketing) reached an inflection point about a year ago. It grew up and realized it had to cast off its childish ways of just “playing around” on social networks and creating pretty pictures on websites and apps. Maturity came with the realization that ROI matters and generating lots of visitors or Fans didn’t translate into the coin in the business’s coffers. Animated GIFs and fancy website features didn’t generate money to pay bills.
Over the last couple of years, firms realized they needed metrics to monitor their digital marketing efforts and make sound decisions that optimized their ROI.
Firms using analytics to make decisions earn 220% higher ROI than firms that don’t
Today, if you’re not assessing your campaigns, you’re likely wasting your time and money!
A host of tools emerged to help firms and Google Analytics introduced new features — today, they introduced content grouping — into their free analytics tool that assesses website analytics.
Tools focused on descriptive analytics — who, what, where, how many. Armed with this data, firms gleefully crafted conversion funnels — like the one above — showing how efficiently they moved consumers down the funnel until some ultimately bought the product or service.
Conversion funnels are fabulous tools showing firms where buyers come from, where visitors drop off, and underscoring problems in the buying process.
But, IS THAT ENOUGH?
Moving beyond Sesame Street
Just like Sesame Street provides the building blocks you need to read, compute, and get along with others, these descriptive analytics are valuable. But, knowing the letters of the alphabet and counting aren’t enough to succeed in today’s complex world, and descriptive statistics aren’t enough to succeed at digital marketing.
A recent article in Forbes offers a use case of predictive analytics and its impact on ROI for Mindjet. This graphic shows the process of collecting and analyzing data to score leads that optimized sales contacts.
Lead scoring is just 1 example of predictive analytics — if you’re interested, I developed a proprietary algorithm that works great for lead scoring.
What makes predictive analytics different?
Rather than just describing the who, what, where, when of your social media campaigns, predictive analytics PREDICT which actions will generate the best return based on an algorithm, such as a regression equation. As portrayed in the image, predictive analytics requires collecting existing (descriptive) data using tools like Marketo, Salesforce, Eloqua, and Oracle. Data is run through a tool like Lattice, SPSS, or SAS to find patterns (correlations/ covariance) in the data. Alternatively, my proprietary algorithm uses an econometric model to build out a sophisticated predictive model. I’ve done the same thing in building my 4-factor model of social media conversion.
Why predictive analytics are useful?
Predictive analytics not only describe what’s happening, they predict what WILL happen in the future, which is REALLY valuable stuff. Here are just a few things you might want to predict:
- Lead scoring — which leads are most likely to convert
- Consumer behaviors — if you do X, consumers are likely to do Y
- CLV (Customer Lifetime Value) — which assesses how AGGRESSIVELY you should pursue a group of customers or prospects
- The optimal frequency for posting
- Optimal pricing
- Customer defection – which customers are most likely to defect
How to build predictive models?
I commonly use a variety of tools. Normally, I’ll start with SPSS to determine the data quality — remove “bad” data, check normality, etc. If necessary, I’ll transform the data (for instance mean centering) and explore the relationships using correlation analysis. If I have a hypothetical model, I’ll test it with the data, if not, I’ll build a model with 1/2 of the data, then validate the model with the other 1/2 — called a jackknife approach.
I use regression (to build linear relationships), cluster (great for finding groups or segments of consumers), and sometimes analysis of variance to understand how groups behave different.y
Implementing a predictive analytics solution
Step 1. Identify your business problem
Step 2. Determine what metrics are necessary to address your problem
Step 3: Determine which analysis technique you’ll use (determines the amount of data necessary)
Step 4: Collect historical data on all necessary metrics
Step 5: Analyze the data including assessments of data quality
Step 6: Communicate findings to organizational decision-makers
Step 7: Implement decisions based on findings
How businesses use predictive analytics
Businesses use predictive analytics in a number of ways, such as the one discussed above. In addition, a number of tools, such as CRM (Customer Relationship Management) use predictive analytics to determine marketing strategies. Another type of predictive analytic is CLV (Customer Lifetime Value) which uses purchase information to classify customers into groups and determine the level of profit reflected by each group, which is used to build marketing strategies for each group.
Descriptive models and predictive analytics
Descriptive models are often overlooked as tools for generating predictive analytics because they suggest strategies that will generate better results without being able to quantify how much better the results will be.
An example is the TRA (Theory of Reasoned Action). This model states that buying behavior is impacted by a consumer’s attitudes and beliefs about the products, as well as the norms related to that purchase. This theory, of course, underpins how social media works. Social media helps build attitudes toward products based on the most credible sources — our friends — and establishes norms of behavior when we see all our friends buying the product.
So, why aren’t these descriptive models used more frequently in businesses? In part, that’s due to poor exchange between businesses and academics who seem to speak different languages.
Predictive analytics and social media
Marketing in general, and social media marketing in particular, are not heavily influenced by predictive analytics. Although, that’s changing as supercomputers allow organizations to use massive data captured during transactions to build predictive models of what consumers buy and factors that impact their purchases.
Still, relatively few companies use predictive analytics to drive marketing strategy. Sometimes, when I pitch to prospective clients, I’m shocked at how few firms demand any true analytics from their agencies and almost none even understand the concept of predictive analytics. If the agency provides any analytics, it’s commonly simple ones such as # of Fans, # or RT, or other somewhat meaningless data.
Agencies and in-house marketing employees often develop simple correlations as a way to build a social media marketing strategy. For instance, they might notice that certain types of content drive more engagement or that posting at certain times generates more engagement, so they do more of this. But, this lacks the depth of understanding necessary to build predictive analytics.
I have several proprietary predictive analytics tools I use to help clients optimize their ROI. For instance, I have a complex algorithm (model) to help businesses generate leads for the sales force from their email marketing programs. In other cases, I build predictive models from scratch or use descriptive models to generate predictive analytics, such as my hierarchy of effects in social media.
Need marketing help to support business growth?
We welcome the opportunity to show you how we can make your marketing SIZZLE with our data-driven, results-oriented marketing strategies. Sign up for our FREE newsletter, get our FREE guide to creating an awesome website, or contact us for more information on hiring us.
Hausman and Associates, the publisher of MKT Maven, is a full-service marketing agency operating at the intersection of marketing and digital media. Check out our full range of services.