Blog

A great article published in the online version of the Harvard Business Review, “7 Strategy Questions: A Simple Approach for Better Execution,” outlines 7 strategy questions you should be constantly asking yourself and those around you about your business.

Here is a short summary of the 7 Strategy Questions mentioned in the article:

1. Who Is Your Primary Customer?

2. How Do Your Core Values Prioritize Shareholders, Employees, and Customers?

3. What Critical Performance Variables Are You Tracking?

4. What Strategic Boundaries Have You Set?

5. How Are You Generating Creative Tension?

6. How Committed Are Your Employees to Helping Each Other?

7. What Strategic Uncertainties Keep You Awake at Night?

This list resonates with those of us here at MindEcology. Each of these questions can be answered – in part or in whole – by applying the right data analytics, data mining, and/or predictive modeling strategies.

Find more a detailed explanation of each at: Harvard Business School Working Knowledge.

No Comments | Permanlink

Many business managers put off making important business decisions – or they make such decisions much too rashly – simply because they feel they lack the sufficient facts upon which to base a reasoned or calculated decision.

After all, there are a lot of intangibles in business. For example, whereas it is fairly easy to measure things like past sales revenue, units sold, or website visits, it is much trickier to measure things like:

* how satisfied are your customers?
* how likely is your new advertising campaign to result in new business?
* how much time should your salespeople ideally spend on the phone with new prospects before moving on to the next?
* how many customers will you acquire next month? next year?

According to Douglas W. Hubbard, the author of ““How To Measure Anything: Finding the Value of Intangibles in Business, you can truly measure anything – often by resorting to making educated guesses or estimates. To measure business intangibles that affect your most important decisions, one of the most powerful tools you have is your own experience (and that of your trusted associates).

For example, think of an important business decision you face but for which you feel you lack sufficient data. Be sure to pose the question in terms of a range of possible values, using this pattern:

“What is the lower limit and upper limit of [intangible], with a 90% chance of certainty.”

For example, you may say:

“I am 90% certain that my company will acquire between 12 and 25 customers next year.”

Initially, most people who are inexperienced with making such an estimate are over-confident, thereby assigning lower and upper limit values that result in a range that is too narrow (and therefore, more likely to be incorrect). However, with the proper training (what Hubbard calls “calibration”), you can learn to make such statements with surprising accuracy. The take-away here is: you know more about your business than you give yourself credit for.

Rather than putting off the major business decisions you are currently facing merely because you do not have access to the hard data you need in order to answer them, trying this handy technique. You will suddenly find yourself much more empowered, and much better-informed, as you move forward in facing a (less) uncertain business future.

No Comments | Permanlink

This month, a California judge ruled that retail firms are no longer allowed to request a credit card holder’s zip code at the purchase point. The plaintiff’s argument was that a customer’s zip code information is not pertinent to the credit card transaction, so asking for it merely amounts to a violation of the customer’s privacy. Here’s the story:

California Credit Zip Code Ruling

Of course, this lawsuit reflects consumer-rights advocates’ ongoing concerns about consumer privacy issues. While privacy is an extremely important concern that deserves ongoing study, this particular battle of the privacy wars was fought on the wrong grounds. The truth is, database marketers already have a veritable treasure trove of information about each and every consumer in the marketplace who has lived at the same address for more than a few months. The data available on all of us goes well beyond the zip code – so much so that most privacy advocates would probably sleep better at night not knowing what marketers already know about them, their neighbors – all of us.

From a B2C marketer’s perspective, knowing more about prospective targets of a planned marketing campaign actually allows everyone to win: the more targeted an individual contact, piece of collateral, or ad is, the better the chances that those who see or hear it will be open to hearing about how it can make their lives better. In other words: more knowledge on the part of the marketer means more precise targeting – which means a potentially happier prospect who has to sort through less “junk” to get to stuff they may actually care about. And, of course, the seller/purveyor of the marketing information is better off because she is focusing her marketing budget on those prospects who are 3-10 times more likely to respond than the average prospect. Spending less money on marketing and advertising can translate to bigger profits for the seller, a lower price for the buyer, or both. Truly win-win.

Ultimately, consumer privacy is an important issue that deserves ongoing debate and attention. But, court battles such as the one highlighted above show that the average consumer does not have the faintest conception of how far marketing has come in terms of depth of knowledge about end users that marketers already possess about them. And, most also fail to understand how targeted marketing can benefit everyone involved in the marketing equation.

No Comments | Permanlink

The process of data mining refers, in general, to the act of applying various mathematical and statistical techniques to massive amounts of often disparate data in the hope of teasing out significant details. These details can then be leveraged for future (usually business or scientific) endeavors.

To make data mining work, one requires access to historical (company) data. The process involves looking for patterns among various independent pieces of data. The results can be leveraged to make better decisions in the future.

As a field, data mining is still relatively new. While some of the statistical techniques employed in data mining have been in existence for a century or more, it has only been in the last few decades that we have collectively had access to the kind of computing power required to carry out analyses of massive amounts of data. And, some of the most important data mining techniques themselves have only arisen within the past 15-20 years, such as the training of artificial neural networks.

Thomas C. Redman, author of the informative book “Data Driven” (2008), quotes a past colleague at Bell Labs who said that, “Data do not give up their secrets easily. They have to be tortured to confess.” We agree with this statement wholeheartedly. To do data mining right, it requires the alignment of multiple skill sets, aptitudes, and opportunities including:

* knowing which types of business decisions can be aided by data mining
* having access to adequate amounts of historical data about your company, such as transaction data or customer data
* being able to judge which types of data are eligible for analysis and which should be ignored
* having the ability to clean up dirty data
* knowing which of the many available data mining techniques to apply to the problem at hand
* understanding how to interpret the results of the analysis properly so that they can be applied to making better business decisions

When carried out properly, data mining can surprise you by making your data spills its secrets about your business past in ways that better inform your business future. By trusting the data that data mining can provide, you may be led to new vistas for your business’s future that you hadn’t even noticed before.

No Comments | Permanlink

Studies have shown that for most companies, the data they have lurking in file folders, electronic media, and in the heads of employees is one of their most under-valued assets. If you are not currently leveraging your customer data, order data, production data, sales data and other key types of data, you are missing out on some important opportunities to grow your business.

You may be wondering whether your company really has data on hand – right now – that you could be using to your advantage. The answer is: you bet you do.

Here are 10 examples of data that can potentially be leveraged for better business performance through expert analysis:

1. Order history data currently housed in your online shopping cart, accounting software, or point of sales (POS) system.
2. Customer mailing addresses, e-mail addresses and phone numbers.
3. Total historical spend (revenue) data for each customer.
4. Current warehouse inventory lists.
5. Historical call logs for outbound calls to prospective customers.
6. Sales performance history organized by sales person.
7. Advertising response rate data divided by media channel.
8. Company purchase history data from vendors.
9. Pricing and product specification information pertaining to competitors’ products.
10. Website visitor history data.

This is just a partial list: you likely have 10-20 more types of valuable data lying around your place of business right now. You just need to learn where to look, how to collect it, and how to make sense of it.

Now that you have a better idea of the potential treasure trove of data that you already have at your fingertips, it is time to:

a. collect it into one place
b. organize it
c. clean it up & fill missing data portions
d. analyze it
e. start leveraging it daily to make better business decisions – and bigger profits for your organization

No Comments | Permanlink

Every businessperson makes decisions every day – literally hundreds of them. Most of these are micro-decisions. Micro-decisions are those little split-second decisions we all make throughout the course of our day, such as: whether to backup your computer files now or later, how to word the e-mail to that prospective client, how much cream to add to your coffee, etc.

While the outcome of each micro-decision is not usually very impactful to your business success, when looked at as a group they are very meaningful and important.

In addition to these micro-decisions, we are also sometimes faced with major decisions. These are much fewer and farther between, maybe coming up in our lives only a few times each week or each month. Into this category fall decisions such as: which customers to pursue for new business, where to build a new physical store location, which type of server to buy, or whether to sign a two-year contract with your biggest supplier. In a business context, the outcome of our major decisions can represent thousands or even millions of dollars of either upside profit or loss. In other words, they are a really big deal.

How we approach any decision can be called our decision-making process. And, how we tackle micro-decisions versus how we tackle major decisions should be very different. Often, however, many of us tend to take the same decision-making approach for both types.

The Power of Sub-Conscious Information Processing And Intuition For Making Micro-Decisions

Micro-decisions come at us so fast and are so plentiful throughout the day that most of us perform little or no conscious analysis on them. In fact, to do so would be turn us into incredibly inefficient businesspeople. Rather than performing a conscious analysis, we (correctly) rely upon our sub-conscious awareness – including our intuition – to make these decisions. It all happens so fast that we are often not even aware that we are making these decisions. They just sort of happen.

The fact that most of our actual cognitive processes happen below the radar of our awareness is a well-documented phenomenon. In fact, neuroscientists know that only a tiny fraction of what we know and how we think is actually within our conscious awareness at any given time. And, we also know that intuition is powerful stuff.

Indeed, our intuition is exactly what allows us to make all of these decisions each and every day. We are essentially off-loading the hard stuff to our sub-conscious mental processor. We just face the decision, think about it for a split-second, and await the solution to pop into our awareness. Simple – and it works quite well for making micro-decisions throughout the day. However, for major decisions in business or in life, relying upon intuition alone is usually woefully inadequate. And, the alternative analytical approaches that most of us employ are not much better.

The Danger Of Taking Major Decisions Too Lightly

When it comes to making major decisions, most of us put a bit more effort into the decision-making process than we do when making micro-decisions. For example, we might call up or meet with some colleagues to discuss it. Then, after one or two meetings, a decision is made and everybody moves on. However, in many cases an adequate analysis is never even attempted. Rather, we come to decisions which the Nobel Prize-winning (1978) scientist Herbert Simon terms “satisficing.” Satisficing is simply an concatenation of the word “satisfying” and the word “sufficient.” A satisficing decision is one that seems “good enough” because the decision:

1. satisfies our intuitive grasp of the situation.
2. stands up to a quick, back-of-the-envelope analysis.
3. seems to be the best solution among the two or three possible solutions initially considered.
4. is often similar to what we have always done.
5. is the one that the majority of the involved decision-makers agree with.

Limitations Of Human Thinking

The trouble with satisficing solutions is that they are highly susceptible to the well-known failings of human logic. Scientific research has shown time and again some of the ways that satisficing solutions fail due the limitations on our ability to make complex decisions rationally. This is what Simon calls bounded rationality.

Specifically, when left to relying solely upon our intuition, everyone on the planet is susceptible to each of the following cognitive (thinking) limitations and flaws on a regular basis:

1. Defining the problem too narrowly, inadequately, or incorrectly.
2. Approaching the problem with either/or thinking, meaning that the possible solution set is too small – all possibilities are not considered.
3. The human brain’s inability to crunch large volumes of numbers at once. (A $5 calculator will beat 99.9% of even the most mathematically-gifted humans when performing simple two or three-digit multiplication problems – every time).
4. The tendency to forget certain facts about a problem situation due to personal bias and/or limits inherent in human memory.
5. Being unaware of or not taking the time to effectively structure a problem so that it can be properly examined and solved.
6. Being influenced by emotions or feelings about a topic and/or by person(s) involved in the decision-making process.
7. The inability to recognize how our own implicit assumptions about a problem situation affect the way(s) that the solution is approached.
8. A tendency to be risk-averse, meaning we want to stick with what has been done in the past rather than trying something new because it feels risky to try something new.
9. The desire to start with a hypothesis and trying to prove or disprove it rather than starting with an unbiased look at what the data say about what is going on.
10. We have limited time and are therefore under pressure to get things done quickly.
11. We can get a bit lazy and want to get past the decision as quickly as possible by whatever means necessary.
12. Being influenced by power relationships (i.e., agreeing with our boss because he or she is our boss).
13. Inability or unwillingness to find, access and gather together the data required for the analysis.

Given that we are the smartest species on the planet, why are we at the same time so flawed? The reason is this: the human brain evolved to make decisions on-the-spot – in the jungle, on the prairie, on the farm. That’s what we’re good at naturally. However, it is only for a tiny fraction of our history on the planet that we have really had the need – or ability – to solve more complex problems.

The Value Of Structured Analysis

Each of the human cognitive flaws mentioned above makes it incredibly difficult to adequately and sufficiently tackle our major business (and life) decisions using the same approach that we use to make those multiple micro-decisions that we make every day.

It is these much more complex problems of which major decisions are made, and these are the ones that require much more than the application of mere intuition or back-of-the-envelope analysis. They require structured analysis. In other words, we need to devote the sufficient time and resources needed to take a structured approach to our problem-solving. Performing a structured analysis provides the following benefits:

1. Allows us to transcend – or at least recognize – our own inherent biases about the problem situation and about possible solutions.
2. Provides the opportunity for intuition to come into play in formulating possible solutions – while at the same time not allowing intuition to dominate the decision-making process.
3. Gives us the ability to mathematically combine multiple data points or variables into a single number (as a ranking, score, or other easily-evaluated data point). It is much easier to make a decision based upon a single score than try to evaluate the merits of 2, 3 or more variables about each potential solution at once.
4. Forces us to take the time to properly state the problem at hand.
5. Gives us the opportunity to see “what is” before deciding what to do.
6. Provides data-driven, third-party-verifiable documentation which clearly highlight our thinking. Such documentation can be shown to other people who are involved in the decision-making process – as well as to other stakeholders such as clients, partners, investors, and suppliers. The documentation can help us to get buy-in on behalf of those not directly involved in the analysis, moving things forward more quickly and with more mutual agreement on the part of all parties.

The next time you face a major business decision whose outcome will impact your business future in important ways, consider investing the sufficient resources to create a structured analysis. Do this instead of relying upon back-of-the-envelope calculations, ad hoc meetings and intuition to guide you to the right answer.

No Comments | Permanlink

Predictive modeling is all about leveraging historical business data to predict the likelihood of future success.

But hold on a moment: if this assertion is correct, its very premise seems to go against that now-infamous – yet ubiquitous – cautionary phrase found on advertisements for investment opportunities anywhere they appear in print:

Past performance is not necessarily indicative of future results.

This line of thinking begs the question: how can what we know about the past predict what will happen in the future?

The short answer is that predictive modeling does not actually predict the future; nothing can. However, what proficient predictive modeling does quite well is that it can tell us with an amazing degree of precision how likely a certain possible future is.

To paraphrase the great 20th century systems thinker and polymath Gregory Bateson, the accuracy of a predictive model depends upon knowing which differences make a difference.

Bateson’s Definition Of Information

Bateson, whose book “Steps To An Ecology Of Mind” served as the inspiration for our company name, famously defined information as a “difference which makes a difference.” In its more abstract interpretations, his definition has broad implications in fields ranging from the theory of logical types to learning theory and from information theory to the study of schizophrenia – even to the so-called “mind-body problem” in philosophy.

In the context of predictive modeling, we can loosely apply Bateson’s definition of information in this sense:

The data that we analyze for possible inclusion in a predictive model can only be viewed as information when it makes a difference in improving the predictivity of the model (i.e., how accurately the model can predict a given future business scenario).

Types Of Future Business Scenarios To Which Predictive Modeling Can Apply

There is a huge range of possible future business scenarios to which predictive modeling can be applied, helping business decision-makers to answer such question as:

a. To which households should we send direct mail – and which should we skip over entirely?
b. Where should we build our next store or branch?
c. To which customers should we try to upsell or bundle high-margin products?
d. In which neighborhoods should we implement a billboard, cable TV, or newspaper campaign?
e. Which companies should we aggressively pursue in order to entice them to relocate to our town or city?

A 30,000-Foot View Of The Predictive Modeling Process

Here is a high-level view of how our predictive modeling process works:

1. Define a business decision whose outcome matters to the business owners.
2. Determine a threshold for the outcome that signifies whether it is desirable or not, either as a binary result (e.g., Yes/No, Good/Bad, etc.) or on a continuum (e.g., sales revenue figures).
3. Collect data (i.e., variables) related to past instances of similar business decisions.
4. Pre-test which individual variables seem to correlate with said outcome and cull those variables from the initial variable list into a “finalist” list of variables.
5. Use artificial neural network software or other machine-based learning technique to train a new predictive model. The objective is to find the most ideal combination and weighting of each of the finalist variables in terms of how well it predicts said outcome with the highest-possible degree of accuracy.
6. Feed into the newly-trained predictive model a new set of data that pertains to analogous instances of the same type of business decision to be made in the future, asking the model to predict the likely outcome.

Modeling Helps Determine Which Differences Make A Difference

As per the above 30,000-foot description, our modeling process reflects Bateson’s “difference that makes a difference” at two points: #4 (pretesting) and #5 (model-building).
The pretesting phase helps us to determine which variables make a difference, while the model-building phase tells us how much each variable matters and in combination with which other variables.

This approach brings the abstract definition of a visionary like Bateson into the realm of practical business and economic decision-making.

No Comments | Permanlink

For multi-unit franchise organizations, deciding just where to put their next location for a store, restaurant or service center can be a daunting task. This assertion should especially ring true for organizations that have 20, 50, 100 or more operational locations under their belt and yet they still do not know why some locations turn out to be superstars while others are dogs. How we can better differentiate between the superstars and the dogs at the site selection stage?

There are many ways to approach site selection for new store locations. Every company who makes these decisions in-house has its own, home-grown methodology that makes logical sense and that is certainly founded in rational thought. However, in practice the models are sometimes highly predictive of store success, while other times they fall short.

At MindEcology, our experience in helping clients in multiple industries to select ideal locations for their ongoing expansion efforts has given us a deep understanding of what makes some locations more successful than others.

In our experience, we find that in-house site selection committees are typically biased in the following 3 ways when it comes to ideal site selection:

* they rely too heavily on the quality and condition of the physical building and its immediate surroundings, while largely ignoring the composition of the population of the surrounding neighborhood
* they allow their personal preferences and anecdotal histories about a particular area influence their decisions
* they focus too heavily on a single data variable (e.g., absolute population count, median age, proximity to a major thoroughfare, etc.) as a determinant for projected store success

By contrast, at MindEcology our data-driven approach to site selection helps us avoid these common biases. Our approach is unique in that we:

1. Base our models on historical store performance data (when available)
2. Leverage advanced segmentation methodologies that transcend the use of more traditional, simple demographic factors such as age, income, and ethnicity factors
3. Pre-test 20, 30 or more individual variables for possible correlation to store success before culling our variable list to a more manageable number
4. Combine the most highly correlative (to store success) variables in unique ways using advanced predictive modeling techniques
5. Our models are internally self-validating, greatly reducing your risk of making an investment in an undesirable location
6. Provide a full set of recommendations for ideal store locations within a give market

In our experience, about 75-80% of store’s future success is relative to “who lives there” (in the surrounding neighborhood). Thus, our modeling process focuses heavily on those factors.

Around here, we often repeat the mantra “Trust the data.” Data-driven site selection techniques will trump the bounded rationality and inherent judgment bias of even the smartest of us humans any day. Trust the data and you can be confident that you have chosen well.

We Use Data to Pinpoint Your Hottest Future Locations

No Comments | Permanlink

People are often surprised when we self-ascribed data nerds at MindEcology demonstrate to them just how much useful information is hidden within their database. Our reports and analyses reveal things about their business that are immediate, insightful and actionable. It is a very satisfying feeling to present one of our reports to our customers and watch their faces as they have a series those little “Ah-ha!” moments that we in business value so much.

Like riding a bicycle or playing chess, what we do is simple – once you already know how to do it, that is. But, it can be hard to explain what we do for the first time to friends, prospects and customers who have not yet seen our work in action.

I find that metaphors are useful in explaining pretty much anything. Recently, when trying to think of the best metaphor for what we at MindEcology do, the first thing that came to mind was that of “reading the tea leaves.” However, upon further inspection, while simple and in a way elegant, the “tea leaves” metaphor misses the mark completely. That’s because the processes that we here at MindEcology use are not based upon any personal powers of prognostication, guesswork, having the “sixth sense,” or being otherwise talented in the area of fortune telling. In other words, it’s not based on what we THINK will happen, but rather what we know HAS happened within your business – along with suggestions for how to capitalize on that for the future.

In that light, I think a better metaphor than that of “reading the tea leaves” for what MindEcology does is that of DNA scientist. As you may have been following in the major press lately, you can now submit to DNA analysis companies a small sample of your saliva – along with saliva samples of your spouse and child – and receive back a wealth of information about you, your spouse and your child. In fact, they will supply you with a gene-by-gene account of what percentage of each parent’s DNA contributed to the creation of your child’s DNA. Very cool stuff.

While we are not DNA scientists, you could say that we are gifted with the ability to look at your business’s customer purchase history data (DNA) and tell you things you never thought possible about where you have been and where you should go next.

Here is a sample of the types of information we can tell you after performing an analysis of your customer database. If you give us 1,000 to 2,000 of your customer records (with just customer name, customer address, total customer revenue figures), we can tell you:

a. who your best customers are, organized by “marketing segmentation cluster” type* – in prioritized order of most-to-least valuable

* note: you don’t have to understand a thing about market segmentation to take immediate and full advantage of our findings

b. which segmentation cluster types are most likely to become a customer if marketed to, and, once a customer, which are most likely to spend more money with you (note: as a bonus, we also show you which types of prospects and customers are NOT worth putting resources into pursuing at all)

c. where to find more prospects just like your best customers

d. how to best go after those best prospects with direct response (direct mail, e-mail, doorhangers) and mass-media (TV, radio, newpspaer, billboards) methods, while “skipping over” those who are much less likely to respond

e. a profile of what your best customers look like, including which types of media (magazines, TV shows, radio shows, web sites) they consume, where they shop, and how they spend their leisure time

f. show you where to build a new store, restaurant or service provider business location based upon finding those geographical areas that have high concentrations of your best customers

In short, we build a custom “model” of your best customers based on actual, historical purchase data that you provide to us for analysis. Then, we show you how to apply that model to your market so that you can: a. do more business with your current customers and, b: find and go after more prospects who are the most likely to convert to being your new customers

Like the DNA scientists you have been hearing about lately in the news, at MindEcology we can take what seems like a relatively insignificant “input” and turn it into an amazing amount of actionable, useful information that you can apply to your business right now for better marketing ROI and better business results.

No Comments | Permanlink

At MindEcology, when we engage a new client for an analysis, we always start with a look at their database in its raw state. Depending upon the project, the data could consist of customer names, mailing lists, order history, or product catalogs, for example.

It is the rule, rather than the exception, that most customer databases that our clients bring to us have not been well-maintained. Whether the data we receive has been stored in a POS (point of sale) system, an online shopping cart application, an MS Excel file, or an MS Access or SQL database – a large portion of most companies’ data is “dirty” data.

Dirty data refers to a large data set which contains a possible wide assortment of imperfections, including missing data values, incorrect value types (e.g., text instead of numeric data for a given field), values that are out of the possible range, incomplete records, etc.

In fact, many of our MindEcology clients are almost apologetic when they grant us accesss to their database for the first time. It is almost as if they have invited us over for a seven-course meal but want us to “please excuse our messy house – it’s been a crazy week.”

Given that this is such a common experience for us, we are never surprised when a new client’s database is found to have not been maintained to ideal standards. Some examples of all-too-common database problems include:

* in an address file (or table): address data showing up in the phone number field

* in an order history file: existence of blank (zero-value) revenue fields for some orders

* in a customer name file: first and last names are transposed for some records

* in a product list file: missing product IDs for products that show up in the order history file (etc.)

Fortunately, in mosts cases this is a total non-issue once we get started on a project. That is because, before we actually start analyzing a database for any project, we carry out what’s called “data cleansing” and, when necessary, data imputation.

Data cleansing involves:

1. validating codes (data values) against a list of acceptable values and deleting or fixing each one as necessary 2. deleting very “dirty” records 3. de-duplicating and merging records

Meanwhile, data imputation involves filling in missing values with intuitive data, such as with a reasonable estimate (this is better than leaving the item blank). There are a number of techniques for doing this.

These techniques are industry-standard in the data mining and modeling fields and help ensure that we have good, clean, well-organized data as we begin the analysis phase of a new project.

No Comments | Permanlink