How Framestretching adds clarity


scaleuseimage

This morning on Twitter @Bedtimemath posted the image above with the caption that read: “A natural distribution spotted in the wild! The wear on a weight machine reveals where people place its pin.

Though we experience life using multiple senses-sound, taste, touch,scent and sight, how much do we actually use to understand our place and actions?  Consider this two dimensional picture, the visual more than likely drives the meaning we make. Did the words add any more meaning to what you understood in looking at the picture?

Last Friday, the strategy discussion I lead monthly talked about visual thinking and I realized how readily my business training reduces most if not all of the perceptions available through my five senses down to only one or two.

At work, we use words or numbers and rarely put both together as well as this tweet and accompanying image.  This visual asks you to interpret the worn out paint or coating and recognize a normal distribution , which is a statistical explanation that adds another dimension to our understanding.

Could you plot a graph with the information you obnormal-sampleserve?  How about wear vs. weight?  That’s only two dimensions, with weight values on the x or horizontal access, and wear shown on the Y or vertical access.  It might look like the two dimensional graph on the right.

Does this representation tell you anything more than  the original image? I assigned numeric values to the different amount of chipped coating. Do you connect the current location of the weight on the machine as  170 across and  estimate as I did 20 up?

Both image and graph show, but don’t tell as much as we assume.

Where’s the context? Do you know the amount of time it took for the paint to chip or relative distance between observations, or usage of the machine itself?  We know nothing about the users of the machine, or its location and yet we do don’t we? The rust itself takes time to form and we can infer that more users choose at least 50 pounds, and the most users 90 pounds.

More than meets the eye

Next time you view a two dimensional graph–ask yourself what’s missing?  Try to voice and articulate the context that you’ve assumed. Yes even if you do it alone, as hearing your thoughts activates different processing.  When you do it in the company of others you will be surprised at the differences in your understanding.

I happened to see that DataScope analytics, a Chicago Based Data science firm had posted a request under data science on Reddit.  They asked “What data skills do you wish non-data people you worked with (e.g. managers, PMs, marketing, HR, etc.) have?  The responses on Reddit were quite fascinating.

Personally, I couldn’t help but notice how the replies typified the  constant challenge and struggle that any information or data presents to everyone.  What does the data mean, what is it’s significance?

The same themes arose in the conversation among business people exploring the challenges of visualizing data, in which we quicly recognized that few people see the appearance of data and instinctively look to explore it, versus others whose interest in data are for the sole purposes of confirming what they know.

The questions and process with which anyone approaches data obviously informs how it gets used and thus represented too.  Too often we use only one or two dimensions as suggested by the graph of the interpretation of the photograph.  I created the variable wear based on my interpretation of frequency approximated by the degree of chipped paint.

In contrast, the two dimensions suggest more than they reveal about how the world works.  Economists, for example, plot  supply and demand curves that look like a large X.  Supply being the first line that descends vertically, and demand the second line that Ascends.  The vertical access is Price and the horizontal access quantity.

What else is  assumed in this representation?  Geography? Time?  what about probability and or frequency?  Accuracy or specific observations as in the photograph are not the point of the representation.  It’s merely to create a general understanding of the relationship between price and quantity from two different perspectives.

The economists are exploring  and not predicting behavior, they are merely seeking ot make sense of the world not necessarily profit from it.

Further explorations that southt to clarify the assumptions led to the evolution of the  behavioral economists and  their additional perspective enahances general  understanding  of people’s beyond the one dimensional buyer or seller role and expanded the representation.  The inclusion of additional dimensions of probability also introduce additional complexity in exchange for greater understanding.

My own training as an analyst has led me to begin with exploration, and interrogate rather than merely to extend or convert  the representations. There’s always ore than meets the eye when it comes to understanding what we see.

 

 

 

Finding hidden treasure in plain sight.


Prospecting, mining both are familiars metaphors describing the activities associated with finding and developing  resource rich opportunities.  Rarely  in plain sight for any passerby  to scoop up and gain advantage, prospecting for Gold, other metals or precious gems like diamonds require active and often deep digging capabilities.

Like precious metals or gems, the secret to good business is creating precious assets of intrinsic value. The attributes to value when known for durability and uniqueness, such as a brand, retain  value over time,  predictably generate  cash flow and  become  difficult for competitors to acquire. But the uncertainty of today’s markets and the disruptive threat of new technologies can quickly erode the value of any asset and so growth is essential.  Whether your strategy calls for acquisition or organic growth, either way, the underlying development and prospecting costs need to be contained.

In stories and legends, merely having a treasure map and knowing where to dig doesn’t always lead to happy endings. Technology has certainly helped to mitigate the risks or advance probabilities of success.  Ground and water penetrating radar and detectors   discriminate ferrous and non-ferrous metals pinpoint the site to begin mining and improves the probabilities of a fruitful yield.  The challenges in any mining activity depend not on the power of the technology or in making the dig profitable. Today’s WSJ headline reads Gold hits $1,700.  Absent reliable, hidden treasure maps knowing where to look is an advantage. Returns depend on offsetting the difficulty and risks associated with its extraction and the quality or grade found. Forbes recently summed this up  USAGX’s Denbow: Gold-Mining Companies Face Challenges Finding New Supply.

Prospecting is a perennial challenge for any and every business, and managing the costs is the key to delivering returns.  The current market turmoil has done more than merely  increase investors uncertainty.  for the Risk averse, who have shied away from innovation  or the adventurous  business who has wisely taken pause, I suggest this is a great time to revisit your strategies.  Standing still can prove surprisingly  advantageous if in the process of cleaning house you discover  undervalued or even overlooked assets.  What value does an earlier project, research or failed product launch buried for any number of reasons offer? Lance and Scott Bettencourt of Strategyn write in Harvard Business Review in June 2011 Innovating on the cheap  a series of suggestions on how you can  leverage your existing assets, or rediscover value in surprising places.

Mining existing assets

I suggest a process that may take you a little further.  Consider Google’s Search business and the  underlying value of its algorithms and index.  Maintaining these assets is of critical importance but so too is the value of constant improvement.  Daily, new content and pages added to the internet require Google’s index continuous update.  Including  rich and diverse content such as images, video and sound  files on the internet challenges Google’s index  and algorithm update to accurately rank and deliver the results.  Realized innovations  continuously contribute  to Google’s financial performance and persistent high  market valuation. Even Google however has failures. Research,  experiences of both internal and external users generate additional  assets hidden in plain sight. Actively sharing and reflecting on the meaning of both successes and failures  allow new project teams ready access to key insights that otherwise would be left to lie fallow collecting dust.  If Google continues to draw value  or benefit from their latent assets, can you?

Identifying data or purpose

Frequently, environmental conditions change a variable’s significance.  Strategyn authors talk about unrealized value in products that may have been premature for the market,  experienced formidable technical difficulties or their launch prevented  by high manufacturing costs . Nothing stays constant anymore.  Consumers are always adapting  their preferences to changing circumstances and environmental conditions, and  business are equally forced to adapt.  A variable’s significance in your business model  in one moment may prove insignificant later. Persistently changing conditions is  why its’ important to frequently revisit your tactical plan and forecast models; and occasionally revisit your business model/strategy.

In 1984, Jesse Jackson was the first Black American to run for President.  I was an assistant statistician working for CBS News assigned  to use the exit poll and early returns to create prediction models to track trends in voters behavior. Race became a significant variable , where as before it had not been much of a determinant factor.  To increase accuracy, the forecast model needed to adjust to accommodate and recognize this historic precedent.  Likewise, when I joined Citibank in 1985, the business needed a P &L model for an innovative new offering in four test markets that linked savings and credit products in a relationship.  No one had looked at  interactive product performance before and the experience was a revelation.  The adaptation to existing analysis and risk management tools were instrumental contributors to the explosive growth of Citibank’s  credit card business. The original business proposition  failed to consider that the risk in a bundled loan or relationship,  product was not merely additive but interactive.  Early, controlled testing allowed them to go back to the drawing board armed with new insights and better understanding of the boundaries.

New data is rarely the culprit in a failure; but as things change,  more data enhances interpretation and  provides insights to re-imagine your business.  When you are the largest issuer of credit cards in the world,  accurate risk models  can be built using available billing histories.  In the 1990’s  mountains of itemized purchase or transaction level was left untouched, though its potential value was clear,  there were no clear benefits to justify the monumental costs of analysis. This was a treasure waiting to be mined.  Lacking urgency or absent a competitive threat also minimized the value of uncovering additional insights into consumer’s behavior.

Fast forward just under 25 years and the costs of time and computing resources to sort high volumes of transaction data is trivial and the returns from real-time processing lucrative. Mined transaction data triggers fraud alerts and delivers additional purchase suggestions based on comparison to  individual consumer history and that associated with cohorts, peers or “friends.”  Amazon  demonstrates   mastery in mining  typical  point of sale enhancements and redeems enormous  value from its dual function processing.

Opportunities and technology capable of mining even richer, more complex data eclipses  the significant value accrued from mining transactions.  The potential  value is driving the collection and complex tagging and sorting  of recorded customer service conversations, video capture of consumers shopping or following their daily routine at work or at home or all the places they go  online, key strokes, eye tracking, written comments.  It appears that there are very few domains of human experience and activity that remain a hold out from data capture.  The number of matching and sorting tools, the algorithms and systems also are getting simpler and more widely accessible.  Today, the speed and volume of results Google returns in a general search is far more advanced than credit card billing records I analyzed.  When was the last time you checked out Google’s  specialized search tools or the technology  coming out of their labs?

Returning Power to the People

The  insurmountable challenges are no longer in finding available data, or even privacy. Its ubiquity and increasing open source availability creates an even bigger challenge,  turning the vast amount of real-time data into a durable advantage.  Sunday’s New York Times (August 7, 2011) reported the unusual establishment in Chicago of a team of specialists tasked to help Chicago harness the technology and gamut of rich data the city collects.  Not alone in its efforts, Chicago is  farther ahead of other governments in creating easy interfaces that contribute to the public’s use of  its treasures of recorded and collected data.  Transparency adds more value by increasing the number of analyst reviewing the information, spotting trends or creating applications that simplify the lives of residents.  For example, the free Bus tracker application to let riders and plan their trip better.  It also holds his office more accountable  and increases the opportunity for activism by city residents.

There’s no doubt that power accrues to those who can imaginatively convert  data into both meaningful and doable innovation.

Finding treasures by leveraging connections

Today’s data mining technologies facilitates more than  accountability and activism.  Beyond knowledge of the type and place of available data,  a dedicated commitment to sift and mine the growing mountains of data requires critical analysis and matching skills.  Google does not stand alone in its specialized capabilities, numerous competitors offer diverse and specialized alternative search tools.  Numerous open source tools  make it easy to sort and manipulate any of the open data made available online.  As in prospecting, the tools and ability may narrow the competition and may advance the process. But those systems capable of exploiting and  enhancing anomalies  with supplementary information increase their chances  to uncover intrinsic value and thus create durable advantage .

Innovation results from capabilities to invent but can equally result from abstraction and adaptation.  Most of us at one time or another have come across a person who managed to re-purpose or refashion an object for an alternative use.   For example, the flower bed below.

Between Naps on the Porch eclectic landscape

Don’t merely consider looking at your existing data in its current form, but revisit it with newer analytic capabilities made possible from the numerous open source and proprietary data mining tools rich in functionality.  Consider supplementing your understanding of your assets from the perspective of your final judge, the consumer.  Also consider these sources:

  1. If images are worth a thousand words, spying consumers who refashion or use products for purposes beyond the manufacturer’s original conception can prove inspiring.
  2. Conversations and story are at the core of social media’s power.  The words of mouth, or stories  associated with transmitting and  promoting your business also motivate, inspire and compel employees to higher performance and deliver insights into how your product can be improved.  How often are you  using these to find products  in your inventory or services, that you may over overlooked or underestimated, but  are important to a group of consumers?
  3. Sales Data–Data mining tools can be used to find surprising blips, if you look beyond the blip.  Focus your analysis on the less understood context such as coincident placements or other variables that may not have made it into your database but none the less explain the anomaly.  They may very well be the source of an unrealized opportunity to refashion and reposition products that have trailed in sales.
  4. Last, perhaps you need to apply data mining tools  on your own data collections. The files of failures, tucked into drawers or file cabinets, the product research and or launches that never saw the light of day may call for another look.  After all, consumer preferences are always evolving, but so are your competitors, as well technology that may allow you to overcome previous cost barriers.  For example, oil and gold extraction from very difficult places is now proving economically viable as both these commodities benefit from high market prices.

More reason to harness data mining technologies to jump-start innovation in product marketing, reuse or refashion your assets to generate additional cash flow.

I’d love to hear of your experiences recapturing value in your business by any other routes as well as  suggestions for good tools or tips to improve your data mining or prospecting success.