Should Your Business Buy or Develop Analytics Capability? A look through the lens of the Amazon-Whole Foods deal

“It’s our recognition that if you go to bed as an industrial company, you’re going to wake up as a software company,” declared Jeffrey Immelt, Chairman and CEO of General Electric (GE) at a Gartner conference back in 2015. He was trying to note that companies should seek to disrupt themselves, making a nod to his own, GE.

Well, he certainly did not imagine making a nod to Amazon, but that comment is so apt given what Amazon did with its bid on  Whole Foods Markets.

Amazon’s acquisition of Whole Foods Markets represents an unprecedented technological investment in the grocery retail industry.  The milestone also signals an unprecedented moment for analytics being seen as a catalyst for long-term business value.

Whole Food Markets Chicago
This Whole Foods Market in Chicago is bustling with customers. The Amazon acquisition promises to help the chain refine its customer relationship through analytics.
A look at an industry and its margins

Grocers like Krogers and Safeway have sold food items and associated services with thin profit margins.  This meant goods would have to be sold at high volume for a grocery chain to be profitable.

When Whole Foods Market started, however, it developed and maintained a margin advantage over other grocery chains through its focus on offering organic foods.  Organic foods command higher prices; thus a high product margin is possible.  Whole Foods Market established a strategic advantage with its organic food focus.  In fact Whole Foods Market is the first grocer to certify its food as organic according to the Wikipedia page on the company.

In recent years, Wal-Mart and Target have begun to offer grocery, increasing competition against traditional grocery chains.  But the most striking tactic has been to offer organic foods alongside those .  Wal-Mart and other low-cost grocers drawing customers seeking bargains away from Whole Foods.

The plan has worked.  Whole Foods reported six quarters of declining sales as a result. Moreover, Whole Foods gained a consumer nick-name “Whole Paycheck”   that references the price concerns people have about its produce.

How Analytics Becomes Essential To Business Growth
Enter Amazon and its potential to give Whole Foods a leg up.

Amazon has a number of analytic solutions that benefit Amazon Web Services (AWS), a set of cloud and services for databases and hosting.  These cloud services have been geared towards either supporting Amazon services or for specific needs of the tech community.  But with the Whole Food acquisition, Amazon has an opportunity to demonstrate tangent value to a industry that consumers and Wall Street analysts naturally understand.  Consumers can see more personalized marketing, which will boost sales.

An effort to provide more value may even challenge that consumer nick-name “Whole Paycheck” nickname.

The Amazon-Whole Food deal creates a transformation for a retail business model: a retailer that has been regularly limited by margins suddenly can determine where to focus on growing customer experience and ultimately retain customers.  It can do so through applied machine learning.  A simple concept behind machine learning is applying computation software to a dataset and providing meaningful results such as personalized service for customers.  A complimentary predictive analytics model can be built to adjust operational support for that personalization service.  Ultimately Amazon can leverage discoveries from machine learning initiatives into meaningful customer and business value.

This scenario is an analytical environment that can benefit Whole Foods against fierce competitors like Target and Wal-Mart, which are already using analytic initiatives of their own.

Decide whether to build software or acquire another company’s 

The decision to acquire a company is no small feat – Amazon’s all-cash acquisition is its largest to date.   But this news places a strategy spotlight on the value of analytics.

Amazon’s growth, albeit with debate about its scale at its start, reflects a trend that scaling a business profitably is part of business growth.  Business leaders are beginning to understand analytics can become a factor for unlocking long term value.  More importantly, they are putting that understanding into action.

The growing appreciation also raises the question – should a business look to acquire a business with analytics capability or should it develop its own in-house analytic services? This dilemma is a significant twist from a classic business question – to develop in-house capability or to integrate another company capabilities that has been honed and refined.

So should companies build software as a response to competitive pressure, or should they acquire it?  There is no panacea for an answer.

Many business owners believe business growth means increasing in physical size.  But doing so increases the demand for capital, and can start a never-ending spiral to induce volume sales.

Analytics can provide the right approach to scale, a variation on Immelt’s thought regarding self-disruption.  But it can take a large costly and timely effort to implement the right solutions.

But no matter what choice managers make, companies have to analyze their strategic activity and determine how to best create solutions that customers appreciate.

Analytics Tips: Basic Data Security for Small Businesses

Clicking A Laptop Computer Keyboard
Keeping data secure starts with managing its sources – although these days an Amazon Alexa is just as likely a data source as a keyboard.

With increased amounts of available data collection comes a variety of new ideas to use data to benefit customers and businesses. An often overlooked consideration is the need to be responsible with the gathered data.

The advancement of technology has increased the inadvertent likelihood of releasing Personal Identifiable Information (PII). PII, as defined in Wikipedia, is metadata that can identify, contact, or locate a single person, or identify an individual in context. Examples beyond a person’s name include email address, date of birth, passport number, vehicle registration plate, and driver’s license.

One kind of PII, an IP address, is not PII by itself, but can be considered as critical PII if it is linked to another piece of data. This impact from blending data means businesses must know not just the meta data collected, but how metadata is combined when used. Without care personal identifiable information can become the property of identity thieves, damaging a company’s reputation as well as lives.

Adding to the concern is public scrutiny of technology while its usage and capabilities are evolving agnostic of consequences. Technology can provide good or bad consequences, depending on its application. For example, pop up offers use Javascript code tailored to the cookie session, but links in a pop up offer can also lead to a nefarious site.

So what should a small business do in this challenging environment?

If you run a small business, a few tips about data can some basics a small business can take heed to some basic informations and be aware of how to act.

Audit Data Usage

Auditing data usage within the business reveals how information flows to critical activity — which systems or employees are used regularly, and if so, what analysis is conducted. Employees and processes should be mapped against opportunities that can potentially lead to unintended exposure, such as unintended data access for employees leave the company and removal of outdated data. Ensure that people who no longer should have access to analytic reports are removed. Another useful effort is to audit data relationships where possible. Neo4j, an open source tools, is good example. You can read more about it in the DMN Tech post.

You can also audit how site elements are called each time your website is loaded. A web proxy or “Packet sniffers” such as Charles and Fiddler allow users to view how each site or app elements are loaded into the browser. These tools can also imply where hack attempts have potentially impact site or app performance, slowing down elements.

One bonus tip: Keep an analytic report filtered to the IP addresses of store locations and branch offices. Doing so can help highlight traffic from potential fraud sources.

Establish Data Guidelines and Removal Policy

Establishing guidelines for managing the storage and retrieval of information can encourage employees to level set to agreed procedures minimize shadow copies of information, which can lead to loss data, theft, and miscalculations.

For example, verifying active accounts on email lists can not only eliminate dead email addresses but also detect email addresses which should not receive data and reports. If you are incorporating web analytics tags with monthly email, coordinate a verification of opt-outs when analytic data is regularly reported. The timing of this verification serves as a reminder.

Share Data Policy with Clients and Partners

Your technology and processes keep data secure, but its your policy that establishes how data is used. Let your customers and clients know a policy is in place so that they understand what your business does to protect their information.

When information is requested, make sure there is an opt out procedure for site visitors. They typically seek statements that they are not locked into one vendor. You can also remind them of opt outs for online racking as a convenience. A low percentage of online users exercise the option, but do want the option. Firefox, Chrome, and IE browsers contain a tracking opt-out for users.

Monitor Tech Media Sources for Privacy Regulation News and Assess Its Impact to Your Guidelines

Finally small businesses can follow news from associations wrangling with the downstream impact of legislation as it is considered. Doing so keeps business leaders informed on what impact legislation can have on their operations. For example, the Digital Analytics Association a “Code of Ethics” for analytic practitioners to pledge. The Code of Ethics is a seven point outline of data stewardship meant to establish a working guideline for an organization’s data usage. The pledge was created in 2011 as a response to controversial legislation that contained a wide interpretation of acceptable tracking solutions and could significantly impact digital agencies and corporate marketing departments alike. The most active associations center around media, while the FTC has gradually raised its scrutiny of online activity.

Data security is not only important to data integrity but to business integrity as well. Developing the right processes that match your operation will not only see how to best improve your business but also show your customers your capability in being responsible with their information. Take heed to these tips, and you will save yourself operational headaches, costs, and your company integrity.

Video – How #Data from #Water Management Impact Quality of Life – @ChicagoCityData

Barrett Murphy
Barrett Murphy, First Deputy Commissioner for the Department of Water Management at the City of Chicago,  explains water management objectives and how data is impacting infrastructure plans to prevent flooding.


Chicago City Data Users hosted its usual meet-up with an unusual topic; Water. The October gathering  examined  how data can be used to improve quality of life as it relates to water.

Barrett Murphy explained the crux behind water management within an infrastructure.  He noted that “Water finds its level”, meaning that it gathers at the lowest point of terrain. In an urban area that tends towards filling basements. Therefore much of the rainwater planning is meant to guide water along the streets to drainage as much as possible.

Murphy walked the audience through many of the obstacles currently facing Chicago’s design. One plan that was discussed was directing emergency water onto the Dan Ryan, which raised laughter in the room as much as it did debate.

Other talks included a GitHub project that monitors beach quality.  Scott Beslow, a Chicago beach enthusiast, founded Drek Beach .  He developed the scoring model on E Coli and included public data from the Chicago Park District .  He has opened the project to volunteers to help refine the model further.

On the site Beslow says “As a civic hacker, open government advocate, and beach lover, I wanted to create a website dedicated to exploring the resulting data.”

Scott Beslow Chicago City Data
Scott Beslow of Drek Beach explains how the site developed. It provides a quality rating for the Lake Michigan beaches in Chicago.

Analytics Tips: Choosing Reporting Tools for Data Visualization

If you’re in business and you have a little data – ok, a lot of data for many businesses –  you may be feeling some fear.  It may seemingly be fear from imagining the preparation that is needed with big data.
But more than likely it’s not.
It’s visualizing the data – or more precisely, how to best visualize it to take action.
Data accuracy is important to build trust not only in analytic solutions but in the people who use the reports from those solutions.  This means marketers must select report interfaces that control the range, look, and feel of the presented information with respect to organizational needs.
In short, what graphic formats should appear with the data on hand?
Let’s look at a few basic options, with a general eye for war they are good for and what drawbacks can occur.
Data Visualization Quality
Take note of what makes the best way to display data visually.
Overall, to use data visualization successfully, managers should consider three aspects in selecting a data visualization platform:
  • Audience
  • Purpose
  • What visually works best for the data
The last point sounds subjective, but selecting the right visual graph is really about assessing trends and deviations. What data would you like to see as a trend, and what would provide a useful alert as a deviation?   A trend reflects what behavior is developing consistently, while a deviation notes a striking change in a trend.  The right graph will help highlight both easily detectable trends and deviations
Spreadsheets Are Better Than Before But Still Have Limitations

Spreadsheets have their value, but they usually offer value for relative small batches of data.  As more columns and rows of data are added, the ability to Increasingly difficult to process relationship among the data becomes difficult.  Spreadsheets present visuals which only allow for isolating figures.  People can’t visually register but one or two numbers at a time.  This is why many plugins use spreadsheets to highlight a change in value.

Keep in mind, however, that dynamic changes may be harder to spot. Use a bar chart or trend chart to help highlight spikes. But using these in a sheet means data updates have to be timely to make the deviations and trends helpful to business decisions.
Templates can provide a starting point…but in some cases just a starting point
Template solutions can provide a standard starting point for viewing how data should be best displayed.  There are a few excellent tools that can visualize data within dashboard options. The most popular tools are usually the self-service BI tools such as Spotfire and Tableau.
There is one drawback with template dashboard solutions: the visualization and concepts are delivered to the user by the solution provider, rather than being custom for certain instances.  A template-based data visualization tool assumes a preconceived notion of what data should look like, or what relationships should exist.
Intelligence from data is not always obvious. This means assumptions from a template-based data visualization tool may not reveal intelligence relevant to a business objective.
What Open Source Data Visualization Provides
Where spreadsheets provides a singular representation of data and basic graphs, data visualization based on open source programming can represent active interrelationships that are constantly updated according to real time data input. This arrangement provides faster reporting and faster responses to the reporting results.
Open source programming solutions are increasingly being repented through Javascript frameworks, such as D3 (, a Javascript library designed for data visualization needs.
Business managers should keep in mind, however, that custom build visualization can sometimes involve a development team which can not only build the dashboard but can highlight the organization’s capability of responding to reporting.  The good news is that because Javascript is being increasingly used, businesses can leverage its web development team.  And in some cases, programming language may not be extensive used, if at all.  Overall managers must first access an organization’s development capability to build an open source visualization that enhances how an organization functions.

#R Programming vs #Python – Infographic via DataCamp

Ready to be a data scientist but nor sure where to start? This useful image describes how R programming and Python are used within data science, from their unique development history to salary commanded in the data science field.   You can compare the pros and cons for each programming language, and learn to select the right one for your development needs.

Infographic courtesy of DataCamp

R Programming Python Data Science