Data Visualization: Providing Data Insights that Cause Change

the internet economy
Courtesy of BCG

Great data insights don’t mean much if the folks controlling change don’t understand them or don’t have the time to pour over columns of data. Enter data visualization; the key to getting data insights to cause change that improves your market performance.

Big data/ big headache

Big data isn’t a new thing; it’s the size of data that’s new. In just the last 2 years, we’ve generated a zetabyte of data from scanners, mobile devices, web connected devices, television, and computers.  To put that in perspective, every 2 days we create as much data as we did from the beginning of time until 2008!

And, every 2 years, we create 90% of all the data in existence.

By 2020, experts predict there will be 50X as much data as exists today.

That’s a lot of data.

But, data doesn’t mean much unless you can do something with the data — and here we’re most concerned about using big data to improve market performance.

Unfortunately, the character of big data is that it’s not only BIG, it comes at your extremely fast so, unless you have a tool that provides automated insights quickly, you’ve lost all advantage from having the data. You need to go from data collection, to analysis, to visualization, to strategy quickly or risk falling behind.

Here’s what Gary King, professor at Harvard University had to say about the big data revolution:

There is a big data revolution. But it is not the quantity of data that is revolutionary. The big data revolution is that now we can do something with the data. The revolution lies in improved statistical and computational methods, not in the exponential growth of storage or even computational capacity

Thus, when we talk about the big data revolution, we’re talking more about interpreting that data — putting it to practical use — than the amount of data generated, which is pretty uninteresting by comparison.

Data visualization

Data visualization is a little like herding cattle — it’s expensive and time-consuming, but, ultimately, necessary if you want to generate profits from your cows.

Of course, data visualization is only 1 means of corralling big data into something useful. Data analysis using statistical tools to generate descriptive, predictive, and prescriptive data analysis also synthesizes meaning from big data.

Even with data analysis, data visualization makes it easier to see not only descriptive data like height, age, and income, but predictive analytics reflecting the relationships among data, and prescriptive data showing the best alternative solutions.

Data visualization tools

Of course, the fall-back position in data visualization is static tools that produce boring line graphs, pie charts, and bar graphs. Even with bright colors, tools like Excel and SPSS produce pretty boring visualizations.

But, regardless of the tool, data visualization, at its best, should uncover new patterns of relationships not visible to the naked eye. Hence, the key to effective data visualization is the ability to capture patterns and relationships in clean, simple visuals that allow the signal to stand out from noise contained in the data.

Some tools for data visualization involve automated analysis of data as it comes in, while other tools involve manually creating infographics or other data visualizations that provide insights. Today, we’ll focus on automated tools used to analyze big data.

A subset of these data visualization tools are interactive, meaning individual users can customize the visualization to match their needs. For instance, a brand manager may wish data visualizations of a granular nature to observe nuances within the data, while the VP Marketing may wish data visualizations providing overviews across the various brands, with less granularity. Key to interactive data visualization is the ability of users to expand analysis; allowing the VP to deep dive into a piece of data to see more granularity.

If you’re interested in exploring options for data visualization, check out this post for a good list of available tools.

The problem of unstructured data

Unstructured data — words, images, video, and all other non-numeric data — represent 80% of the data available according to IBM and other digital data experts. But, the tools available for data visualization of unstructured data lag far behind those for visualizing quantitative data because the statistical tools necessary to derive meaning from unstructured data reflect a similar lag.

Tag clouds or word clouds are a rudimentary type of data visualization that transforms words into a graphic reflecting the frequency of word usage.

Association trees depict relationships

Courtesy of Information Management
Courtesy of Information Management

among words used. For instance here’s an association tree from Information Management showing relationships between customer sentiment about a brand:

Cubism Horizongraphs  is a tool for analyzing video and audio files. Using this tool to analyze call center interactions or sales presentations offers insights about customer problems, decision factors, and intentions.

Analysis provides data visualization insights from multi-dimensional data from a variety of sources, thus providing a 360 view of a customer or prospect.

Courtesy of Information Management
Courtesy of Information Management

Network Graphs, like this one, show semantic relationships from large, contextual datasets. In this graph, you see the relationships between the characters in Les Miserables.


Data visualization for unstructured data: Software

New tools for using unstructured data to create data visualizations crop up all the time. Some of the big boys, SAP, IBM, SAS, and others have tools that adapt to unstructured data. In addition, here are some specific tools for handling unstructured data visualizations:

Datawatch provides a solution that creates visualizations from structured and unstructured data and, most importantly, integrate across data types and sources. Options for data visualization includes tables and grids, surface plots, stack graphs, and spread graphs. All data visualizations are interactive and they offer a dashboard for integrating across data. They offer a free trial and training, but don’t display pricing on their website.

Workday Big Data Analytics provides a solutions for data visualization for both structured and unstructured data.

MarkLogic provides a solution, Tableau, for visualizing structured and unstructured data that doesn’t require coding or IT support.

Zoomdata provides data visualization for semi- and unstructured data using tools like Hadoop and NoSQL to create interactive dashboards for visualizing historic or real-time data.

Need help?

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 the 1st chapter of our book on digital marketing analytics – FREE, or contact us for more information on hiring us.

Hausman and Associates, the publisher of Hausman Marketing Letter, is a full service marketing agency operating at the intersection of marketing and digital media.

You might also like:

An Insider’s Guide to Digital Marketing Analytics

digital marketing analytics
Courtesy of Avinash Kaushik

Forbes proclaimed 2014 the Year of Digital Marketing Analytics, summing up the problem this way:

If most digital marketing programs or campaigns have a weak area, it’s analytics. One recent study identified that the biggest talent and hiring gap in online marketing is in the analytics space. 37% of companies surveyed said that they desperately needed staff with serious data chops.

If you take a look at the image above, courtesy of Avinash Kaushik on Occam’s Razor, you’ll see a similar emphasis on “Big brains” and there just aren’t enough of them going around.

The state of digital marketing analytics today

Well, in 2015 we still find too few analysts trained in digital marketing analytics, especially when it comes to more advanced analytics. What passes for digital marketing analytics is also pretty dismal, amounting to little more than rudimentary vanity metrics.

If you look at interest in digital marketing analytics over time, you find the term first appeared in search in 2011, but searches exploded in 2013. Google forecasts continued steep growth in searches for digital marketing analytics based on the graph below from Google Trends.
trends in digital marketing analytics

So, what do these searches turn up?

A ton of tools, many of which aren’t really analytics tools, but automation tools with a little tracking. For instance, I love SproutSocial for helping share and curate content, but it’s not really an analytics tool. Here’s what you get:

Reports | Sprout Social 2015-01-16 09-55-43

I ask you, how does this data help manage your digital marketing? What insights does it provide?
The same goes for many “analytics” tools provided by the social networks, which are pitifully anemic. A couple of caveats here, however. Google Analytics and Facebook Ads Manager provide very useful, insightful data to help optimize your digital marketing results. I’ve provided detailed directions for setting up and interpreting data from Google Analytics and Facebook’s Ads Manager.

What you need to rock digital marketing analytics?

Surprisingly, the first step is to gain an appreciation of analytics. I find many small and mid-sized companies don’t appreciate how critical digital marketing analytics are for their success. Even some large businesses don’t really get the importance of digital marketing analytics and focus too much on late funnel assessment rather than top of funnel assessments.

Recognize that digital marketing analytics require a budget. Too many businesses try to go cheap here with the notion that money is better spent on other activities. And, in the short run that might be true. Unfortunately, what you’re not seeing in this cost strategy is the opportunity cost of sales you didn’t make because your efforts weren’t optimized. I call this a penny-wise and pound foolish strategy because you’re saving a little money up front to lose a lot of money on the back-end.

kpi and metricsKPIs and ratios

Next, you need to build KPIs (Key Performance Indicators) and metrics from your mission and strategy, focusing on both top of funnel (consumer sentiment, reach, engagement) and bottom of funnel (ROI, conversion, etc) strategies. This is why you need marketers schooled in digital marketing analytics — they understand marketing KPIs.

Set realistic priorities because you can’t focus on every possible KPI at the same time. I recommend selecting a balance between the KPIs at the top, middle, and bottom of the funnel that have the greatest impact on market performance.

Setting goals for these KPIs allows you to develop more meaningful metrics like ratios of expected versus actual. Large ratios demand investigation (and maybe testing to figure out why the ratio was large) while small ratios indicate you met expectations.

Level of analysis

Also, think about level of analysis issues — you want both overviews of how well your strategy is working and insights into segments, such as different social platform performance, performance of different types of content, etc. As an analyst, think about what different users need in terms of level of analysis. For instance, the VP marketing needs an overview, but she might want to deep dive into why some KPIs had high ratios. Meanwhile, your brand managers want to understand the performance of their products and community managers the performance of individual pieces of content. These elements fit within Kaushik’s notion of dimensions that covers performance of individual keywords, campaigns, posts, referring sites, countries, types of visitor, etc.

Data visualization

Visualizing data is critical for easing interpretation. In his TED talk, David McCandless, said this about the importance of data visualization:

By visualizing information, we turn it into a landscape that you can explore with your eyes, a sort of information map. And when you’re lost in information, an information map is kind of useful.

Data visualization not only acts as a short cut for interpreting data, the human eye sees pictures a whole lot better than numbers. Thus, appropriate visualizations allow managers to identify problems quickly so they can fix the problem before it becomes a crisis.

For instance, P&G monitors deliveries using GPS installed in its fleet of trucks using colored digital block — each block representing the value of the customer to P&G and the color representing expected delivery (green for on time, yellow for possible delays, and red for likely delays). When a truck runs into problems (traffic, weather) that threaten delay to a major customer (like Wal-Mart), managers can quickly send replacement shipments from a local distribution center or re-direct shipments from less critical customers or shipments with sufficient lead time.

Translating digital marketing analytics into action

Unfortunately, many firms find their digital marketing analytics programs falling down at this critical step — translating insights into action. In this article, Google quotes poet, Andrew Lang who eloquently said:

He uses statistics as a drunken man uses lampposts—for support rather than illumination

Translating insights into action often means going back to manipulate your data for more nuanced insights;

  • looking for relationships among your data – for instance, you might uncover a relationship between top performing posts and specific keywords used or publication timing
  • looking at trends rather than data points – trends often help you identify meaning in your data such as cyclical trends or when a particular data point stands out from others versus simply representing normal fluctuation
  • turn data into predictive models – don’t stop with viewing data as isolated points and basing forecasts on simple linear extrapolations. Predictive models use historical data to determine the relationship among a set of factors and desired outcomes (like KPIs). Then analysts use these algorithms to predict future KPI performance. You can even play “what-if” games to determine the impact on performance of various actions. This helps determine which changes represent the greatest impact on performance.
  • don’t forget that data analysis is part art and part science. Translating insights into action involves a certain amount of playfulness with the data to discover deeper insights.

Need help?

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 the 1st chapter of our book – FREE, or contact us for more information on hiring us.

Hausman and Associates, the publisher of Hausman Marketing Letter, is a full service marketing agency operating at the intersection of marketing and social media.


You might also like:

Tips for Measuring the ROI of Digital Marketing

social analytics suckIn the bad old days, digital marketing was a free-for-all where instant gurus touted their money-making formulas (usually little better than snake oil salesmen) and deluded followers into spending thousands for coaching programs that didn’t work. Of course, without metrics for measuring the ROI of digital marketing, these gurus continued raking in the money from gullible and desperate businesses.

This isn’t a new problem and it’s unique to digital marketing. As far back as the late 1800’s John Wannamaker is quoted as saying:

Half the money I spend on advertising is wasted; the trouble is I don’t know which half.

Traditional advertising faces a similar problem with companies allocating 60% of their media budget to television when only 18% of TV advertising campaigns generate a positive ROI, according to Nielsen.

Now, of course, digital marketing is much more sophisticated and it’s harder for false gurus to seduce business owners without proving the ROI resulting from their digital marketing programs. Below are results from studies showing the ROI of digital marketing:

ROI of digital marketing

the ROI of digital marketing

  1. A study by Microsoft used big data to measure the ROI of digital marketing both with and without traditional advertising. They found digital marketing outperforms all forms of traditional advertising (TV, print, radio, and outdoor), while combining both resulted in the highest ROI. Thus, digital marketing isn’t an either/ or strategy, but businesses should blend traditional advertising and new media. Also, businesses whose media spend is still focused on traditional advertising should migrate their budgets in favor of digital marketing.
  2. A case study by Google and Dove showed a 6% lift in sales, while combining traditional advertising (TV) with digital marketing resulted in an 11% increase in sales. Interestingly, the study showed the “tide lifts all boats”. In other words, advertising a single product through digital marketing caused an uplift in sales of other Dove products.
  3. Nielsen showed that CPG (Consumer Packaged Goods Companies) demonstrated the positive ROI of digital marketing was nearly 2.8%, with some industries showing an ROI of over 5% — not too shabby.

The state of ROI assessment

The state of ROI assessment is dismal, according to the Fournaise Marketing Group, which found:

Nine out of ten (90%) global marketers are not trained to calculate return on investment (ROI), and 80% struggle with being able to properly demonstrate to their management the business effectiveness of their spending, campaigns and activities, according to new research.

Why is ROI assessment so bad?

Fournaise CEO identified 2 problems in their study that account for the dismal state of measurement of ROI in digital marketing (or marketing in general, for that matter).

The first is the poor training of marketing majors in assessment of marketing ROI and the second is the influx of non-marketing majors into the marketing discipline (over 1/2 of all marketing employees have non-marketing degrees, most often in the social sciences). He sums up the problem with this statement:

In other words, every Tom, Dick & Harry is a Marketer, lacking the scientific and financial knowledge necessary to inform and optimize the creative side of Marketing. CEOs have told us again and again: they want ROI Marketers, i.e. 360-degree performance machines trained to deliver (real) business results and prove/optimize ROI. As long as Marketers continue to fail to get trained in, master the use of and optimize Marketing Performance & Marketing ROI, they will struggle to demonstrate to CEOs that they are not ‘money spenders who jump on (and hide) behind the latest fads and blow smoke’, but real business generators

ROI of digital marketing and market performance tips

First, let’s take a look at digital marketing and where it fits within the spectrum of traditional marketing. Here’s a very cool infographic I created with the help of Matt Valvano from Ideas and Pixels — a first-rate graphic designer.

digital marketing strategy

The infographic shows the various elements necessary to achieve positive ROI of digital marketing campaigns. Basically, 2 things account for positive ROI:

  1. bringing more visitors to your store (or estore)
  2. convert more visitors who show up at your store or estore


Unfortunately, many attempts to measure the ROI of digital media focus on these end results, totally ignoring the variety of factors that generate positive outcomes — a very dangerous practice.

Tip #1: Think beyond outcome measures

So, my first power tip for measuring the ROI of digital marketing is understanding the complex set of activities and interrelationships among activities resulting in positive ROI. For instance, a focus on building a social media community backfires quickly if you have problems with customer satisfaction due to poor product performance — all you’ve done is give disgruntled customers a platform for complaining about your product or service.

Tip #2: Measure what matters, not what’s easy

Often you’ll find digital marketers measuring the easy things — likes, clicks. Sure, these things matter (somewhat), but they’re not the most important (or only) important aspects of a successful digital marketing campaign.

First, set clear goals for your digital marketing campaign — goals that go deeper than just outcome performance measures. Then, create KPIs (key performance indicators) related to those goals.

If you’re convinced customer satisfaction impacts market performance (as is the case for most businesses), assessing sentiment makes a lot of sense. But, don’t stop with sentiment analysis — look at the totality of KPIs and measure all of them. Better yet, chart performance across all KPIs over time, which is much more insightful than putting all your faith in point measures.

Tip #3: Metrics aren’t enough

Don’t simply create dashboards with displaying your metrics. Statistics don’t speak for themselves and require interpretation by skilled analysts combining both the art and science of analytics to uncover actionable insights from your metrics.

While we’re on the topic of dashboards, think about issues related to the level of analysis appropriate for different users. For instance, the VP marketing needs a broad overview of metrics related to the entire product bundle, while brand managers need a more detailed view of just the products they handle.

A good dashboard allows users to dive deeper or take a broader overview of metrics. Also, adding the ability for users to create ad hoc reports and alternative visualizations increases the effectiveness of your dashboard.

 Tip #4: Tie compensation to metrics

One of the biggest challenges firms face (once they get over the hurdle of generating meaningful metrics) is translating data into insights then applying those insights to actions. So, it’s a good idea to tie compensation to metrics — this ensures your employees pay close attention to metrics and try to optimize market performance by using insights provided through these metrics.

I have 3 caveats, however, when it comes to tying compensation to metrics:

  1. Balance the compensation to ensure it’s challenging to achieve higher levels of compensation without being too difficult to achieve. If you expect too high an ROI of digital marketing employees (something unrealistic) they won’t try. If the expectation is too low, they’ll leave money on the table by not doing everything possible to optimize your digital marketing campaigns. You also want to pay attention to the degree to which compensation fluctuates based on performance. There should be adequate incentives to optimize the ROI of digital marketing.
  2. Be very careful that you’re compensating employees for metrics that correlate highly with the ROI of digital marketing. Tying compensation with vanity metrics, like # of Facebook Fans, will drive behavior toward achieving a large Facebook fan-base. However, there’s strong evidence that absolute size of your Facebook community matters little while the engagement of your community provides a stronger impact on the ROI of digital marketing. Pay for what matters.
  3. Employees must have control over factors impacting metrics. For instance, marketers might have little control over customer satisfaction if the production department turns out a really crappy product or logistics can’t get the product delivered in a timely manner. Employees quickly become dissatisfied with a compensation plan containing elements they don’t control.

Tip #5: Don’t stop with descriptive analytics

Move past descriptive analytics (how many, how much, how often) to employ predictive analytics.

In essence. predictive analytics build models using big data to uncover relationships among the factors that impact the ROI of digital marketing (or any other variable of interest).

Your turn

What advice and tips do you have for improving the ROI of digital marketing?

Need help?

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 the 1st chapter of our book – FREE, or contact us for more information on hiring us.

Hausman and Associates, the publisher of Hausman Marketing Letter, is a full service marketing agency operating at the intersection of marketing and social media.

You might also like:

10 Myths About Predictive Analytics

social media analyticsBusiness is tough and competition can be brutal. With many economies still sluggish after the financial Armageddon that caused a mortgage meltdown and the stock market to experience the biggest losses since the Great Depression, it’s never been harder to make a buck. Businesses trying to gain whatever competitive edge they can turned to big data and predictive analytics hoping to “exploit new opportunities and gain the upper hand over competitors” according to TechRadar. In their interview with James Fisher, SAP’s VP of Marketing and Analytics, Fisher talks about the importance of predictive analytics for driving business success:

Predictive analytics technology is the core enabler of big data, allowing businesses to use historical data, combined with customer insight, to predict future events. This could be anything from anticipating customer needs, forecasting wider market trends or managing risk, which in turn offer a competitive advantage, the ability to drive new opportunities and ultimately increase revenue.

No wonder more businesses use predictive analytics today than ever before. Not only is there more data available, but the cost of tools puts them within reach for most businesses [check out predictive analytics solutions by IBM’s SPSS Modeler, SAS Enterprise Miner, SAP Predictive Analytics, and Oracle’s Data Mining ODM]. In the past, you’d need lots of expensive storage, massive computers, and expensive software to run predictive analytics. With the reasonable cost of cloud storage and inexpensive software that runs on the average desktop computer, everyone can harness the power of predictive analytics.

Not true.

It turns out, even though more businesses are trying to make sense of their data, few succeed. Although a recent study by SAP finds more than 85% of surveyed businesses use predictive analytics and 77% believe they’re getting higher revenues because they’re data-driven, it turns out most are leaving a lot of money on the table — at least figuratively.

So, why are businesses large and small failing with predictive analytics?

Maybe they just drank the Koolaid and fell prey to the many myths about predictive analytics.

Now, don’t get me wrong. Anyone who’s read this blog knows I’m a strong advocate of data-driven marketing (and business intelligence, in general). The problem comes when businesses fail to understand how to do predictive analytics the RIGHT way.

Are you falling for any of these 10 myths about predictive analytics?

Let’s see.

Myth #1: Predictive analytics is easy

Sure, new tools, like the ones listed above, make it easy to analyze big data and derive “answers”. In fact, you can throw in data and basically just let the machine run until it spits out something. The problem is, the answers might not be worth the energy it took to make the calculations.

Running analytics programs is easy, doing it right is hard.

That’s because you don’t just turn the computers loose a let them run. Predictive analytics requires some serious training in consumer behavior (at least within the marketing area) as well as alignment with company goals.

I remember when I took a class from IBM on using their software. Within a couple of hours we figured out how to run data and the different options for analysis. Then, they turned us loose on a real data set and I remember staring at the screen, not knowing what to do next. That’s because I didn’t have a theory about how the data might be related — a necessary starting point for running predictive analytics effectively.

Myth #2: Scientific evidence is proof

Just because folks say something, doesn’t mean it’s true.

A great example is the New Coke debacle. Coke’s market research folks went out and asked consumers about their preferences for soft drinks. Coke used their responses to develop a formula that better matched preferences.

It failed.

People were in the streets protesting New Coke (called simply Coke).

People hoarded (Old) Coke so stores quickly ran out of their stock.

It was a public relations nightmare — or maybe not considering the $millions in free publicity Coke got.

Why did New Coke fail?

Simple. Folks didn’t just buy it for the taste. They bought it for the whole brand image — the nostalgia of having grown up drinking Coke.

Coke never asked them about that. They never thought about it. They didn’t really understand their customers.

Myth #3: Only what you can measure matters

Predictive analytics relies on metrics — many of them historical data, some from studies. There’s the prevailing notion that things only matter if you can measure them.

But, that’s not the case.

Sometimes things you can’t measure make a whole lot of difference.

Take trust, for example. Does trust impact whether folks buy your stuff? You better believe it.

You might be able to infer trust because someone buys your product, but you can’t directly measure it. So, it doesn’t show up in your predictions.

Myth #4:Correlation = causation

Predictions are primarily based on correlations (relationships) between the data you have.

But, correlations don’t mean that one factor CAUSED the other factor. Just because 2 things are related doesn’t mean one caused the other.

The best example of this is the correlation between hem lengths and the stock market — the shorter women’s skirts, the higher the stock market. But hem lengths don’t cause stocks to go up any more than high stock prices force skirts up. In fact, both are caused by confidence and a sense of well-being. So, if you try to manipulate stock prices by forcing manufacturers to shorten skirt lengths, you’ll fail.

Myth #5: Predictions are perfect

Predictive analytics produce probabilistic estimates of the future. No one has a crystal ball and predicts with complete accuracy. Take horse racing for example. People, knowledgeable people, place bets on horses using predictive factors like age, bloodlines, prior performance … The odds reflect the combined predictions of all betters. Most of the time, the odds on favorite wins — performs as predicted. But, every once in a while, the long shot surprises everyone and takes the purse. This leads to the next myth….

Myth #6: Predictions are forever

Not so, as we saw with our horse race. More data usually makes predictions better. As time goes on, new data should be added into your model and better predictions made about the future.

But, sometimes, the whole model goes haywire. Cultural shifts, demographic changes, and other events might drastically change the model.

Myth #7: You need a skilled consultant to implement predictive analytics

Not so. Go back to Myth #1 and you’ll see the skill necessary to run accurate predictions. But, hiring an outsider might not be the best way to step up your predictive analytics program. Predictive modeling requires an intimate understanding of what data is available or can be collected, the goals of the organization, insights about the organizations culture, structure and market …

Outsiders rarely have the internal knowledge necessary to run an effective predictive analytics program. Instead, you might have to invest in hiring or training employees.

Myth #8: Predictive analytics is mostly a machine problem

Somewhat related to some earlier myths is the notion that predictive analytics is a black box. You pour data in and something happens in the box (computer) that yields accurate predictions. It’s an appealing notion, but not completely accurate. Pouring in data often generates a lot of spurious correlations that don’t really mean there’s a relationship between factors.

That said, sometimes pouring in data and seeing what comes out is an effective FIRST step in predictive analytics. However, the resulting predictions must be validated before a company uses them for planning.

Myth #9: Predictive analytics are expensive

As I mentioned earlier, predictive analytics doesn’t have to break the bank. New software and cloud storage make it within reach of most businesses.

Myth #10: Insights = action

This may be the grandaddy of all myths about predictive analytics.

Predictive analytics, done effectively, produce insights. Turning those insights into action takes both intuition and managerial skill to gain buy-in from stakeholders and pivot the organization.

Your turn

What are your experiences with predictive analytics like?

Do you have other myths to add to my list?

Need help?

We welcome the opportunity to show you how we can make your marketing SIZZLE.  Sign up for our FREE newsletter, get the 1st chapter of our book – FREE, or contact us for more information on hiring us.

Hausman and Associates, the publisher of Hausman Marketing Letter, is a full service marketing firm operating at the intersection of marketing and social media.

You might also like:

Why Predictive Analytics is Better than Stuffing with Turkey

thanksgivingthanksgiving predictive analyticsYum. Maybe I’m just channeling Thanksgiving — which is only a few days away, but this title just jumped into my head as I was working on today’s edition of Analytics in Action. What do predictive analytics and stuffing have in common — probably nothing, but you know the yummy goodness of stuffing with your Thanksgiving turkey, right? It just makes you all warm and fuzzy inside (not to mention sleepy because of all the tryptophan in the turkey).

Well, that’s the awesome goodness you get when doing predictive analytics.

Predictive analytics recipe

Just as everyone has their favorite stuffing recipe, every business has its own recipe for predictive analytics — what to measure.

If you want to learn more about putting together a predictive analytics recipe for your business, I’ve collected some resources for you:

  1. HBR’s Predictive Analytics Primer
  2. Information Week’s Big Data Analytics: Descriptive versus Predictive versus Prescriptive
  3. Forbes: How Big Data Helps Stores Like Macy’s and Kohl’s Track You Like Never Before
  4. Forbes: 5 Steps to Master Big Data and Predictive Analytics in 2014

That’s because businesses have different goals, different available metrics, and different consumer types. So, what do you need for your predictive analytics recipe:


Data is the bread that forms the bulk of your predictive analytics. Without good quality data, your predictive analytics recipe falls short, just like your stuffing isn’t very good without good quality bread.

To predict what customers WILL do, you need to understand what they ARE doing and why they’re doing it. I think an example helps, so here’s my 4 factor model of social media performance:

Social media performance = amplification X sentiment X marketing intensity X close rate

Using this model, a firm gathers metrics for each factor:

  • social media performance (achieving conversion goals like clicks, downloads, purchases)
  • amplification (engagement such as shares, likes, RT … results in amplification)
  • sentiment (how do customers and prospects “feel” about your brand — usually in terms of positive, negative, and neutral emotions)
  • marketing intensity (assesses your digital marketing efforts ie. # posts, $ on digital advertising)
  • close rate (which is pretty self-explanatory – what % of visitors to your site buy (or subscribe, depending on your goals), what percentage of email readers click on a link, etc.

Notice some metrics are actually computed from a variety of sources, rather than a single metric from a report. For instance, to calculate amplification, you’ll need to add up engagement on whatever social media platforms you use plus the shares directly from your website.

We commonly refer to these metrics as descriptive statistics because they describe what occurred. While descriptive statistics are useful, they’re just not enough to optimize your business strategy. Hence, collecting descriptive statistics related to amplification (engagement across social networks) really helps show which times of day and days of the week perform best for creating engagement — especially when you use systematic testing as input for these descriptive stats — they’re not as powerful as using metrics as input for predictive analytics.


Sure, you can build a predictive model directly from the data, especially when you have so-called “big data” through a process of data mining, these models often perform poorly — for more on this stay tuned for my next post.

Data mining is a process of data reduction based on correlations. If you remember your college stat class, you’ll recall that correlation is NOT causation — meaning that just because 2 things change together, they aren’t necessarily related. A perfect example of this is the old adage about the correlation between the economy and hem lengths — a robust economy tends to feature shorter hem lengths, while a poor economy features a longer skirt.

If there were a direct correlation between hem lengths and the economy, we could fix the economy by requiring shorter skirts. Unfortunately, hem lengths don’t cause changes in the economy nor does the economy affect hem lengths. Instead, the popular notion is that both trends result from optimism, which isn’t part of the model.

Anotherspurious correlations problem with data mining are spurious correlations — relationships that don’t have any meaning. Here’s a hilarious example of a spurious correlation (see more here):

Instead, we need to develop inferences based on our understanding of consumer behavior. For instance, my 4 factor model is based on a similar model used to evaluate sales person performance that’s been tested many times. Since sales person performance is somewhat like social media performance, it’s a valuable inference. Testing will show how well the 4 factor model actually predicts social media performance.

For instance, Vera Wang uses a predictive model to send more highly targeted emails resulting in 63% fewer emails, 101% higher CTR (click-through rates), and 275% higher conversion rates. Not bad inferences.


Next, you take your inferences and build statistical models — usually employing regression analysis (or related techniques like logistical regression — logit). Statistical models turn inferences in useful insights to guide strategic planning. Basically, regression uses historical data to assign beta weights to the individual factors (variables) in the model. These weights not only let you know which variables have the greatest impact on your outcome variable (goal), but allow you to predict goal achievement from various strategy scenarios.

Going back to the Vera Wang example, they likely looked at goal achievement (CTR and conversion) across recent emails sent to their email list. They predicted which types of list members responded to which types of messaging. Then, they customized their messaging sending each member the type of message most likely to motivate them to respond positively.

Car dealers use predictive analytics the same way. They know how long owners of particular cars keep their car. Using this data, they purchase vehicle registration data for owners who are nearing the time when they’ll trade in their existing car for a new one. By selectively reaching car owners likely shopping a new car, they reduce costs and increase returns from their marketing efforts.

Sharing insights

A key element in your predictive analytics recipe is sharing insights. Statistics do NO good if they’re stuck with data analysts rather than reaching decision makers, which is where predictive analytics turn into profits. Otherwise, your predictive analytics are just novelties.

Changes to 2014 predictive analytics

If I were writing this post a few years ago, I’d finish without adding this section. Predictive analytics used to depend entirely on quantitative data — numbers. Now, with the explosion of social media, we need to think about how qualitative data — words — predict the future of a company. Even my 4 factor model contains a qualitative component — sentiment. Sentiment comes from coding utterances from forums, social networks, even customer service calls to score the overall sentiment related to your brand.

But, you can drive even better predictions without coding qualitative data — using tools like nVivo and Hyperresearch to create insights directly from qualitative data. For instance, I might use something like cluster analysis to create groups of customers based on similar utterances. Groups who buy my product for gifts or special occasions require a different marketing tactics than consumer groups who buy my product as a routine part of their weekly shopping trip. And, the represent a distinctly different CLV for my business.

Need help?

We welcome the opportunity to show you how we can make your marketing SIZZLE.  Sign up for our FREE newsletter, get the 1st chapter of our book – FREE, or contact us for more information on hiring us.

Hausman and Associates, the publisher of Hausman Marketing Letter, is a full service marketing firm operating at the intersection of marketing and social media


photo credit: Sugar Daze via photopin cc


You might also like: