Data Informed

McKinsey Report: Big Data at Center of Disruptive Technologies

Christopher Nerney — Thu, 23 May 2013 22:51:35 +0000

A dozen potentially disruptive technologies could deliver as much as $33 trillion in economic value around the globe by 2025, according to a massive new report from the McKinsey Global Institute. If you’re expecting to see big data on that list, you might be disappointed; the 12 technologies include the mobile Internet, cloud technology, renewable energy, energy storage and 3D printing, but no big data. A c0-author of the report, Michael Chui, told attendees at the MIT CIO Symposium on May 22 that while "big data underlies all of them in some way or another" the large datasets and analytics tools applied to them are directly relevant to four areas of activity: knowledge work, advanced robotics, next-generation genomics and the Internet of Things.

Related Stories

An introduction to making big data technologies work in the enterprise. Read the story »

Finding analytics value from the Internet of Things means keeping pace with machine data. Read the story »

GE invests in project to embed predictive analytics in industrial Internet. Read the story »

If you’re not keeping track, nine of the 12 categories in the McKinsey report are mentioned above. The remaining three are advanced oil and gas exploration and recovery, autonomous and near-autonomous vehicles, and advanced materials. Here’s how McKinsey tied big data to several of the disruptive technologies detailed in the report: Next-generation genomics. “Next-generation genomics marries advances in the science of sequencing and modifying genetic material with the latest big data analytics capabilities. Today, a human genome can be sequenced in a few hours and for a few thousand dollars, a task that took 13 years and $2.7 billion to accomplish during the Human Genome Project.” Knowledge work automation. “Advances in software, especially machine-learning techniques such as deep learning and neural networks, are key enablers of knowledge work automation. These techniques give computers to ability to draw conclusions from patterns they discern within massive data sets. … Moreover, these machines can ‘learn’ more and get smarter as they go along; the more they process big data, the more refined their algorithms become.” The Internet of Things. “In the oil, metal and mineral extraction industries, Internet of Things technology could help find and map mineral deposits and increase recoverability. Operating costs reductions of 5 percent to 10 percent have been realized through the use of sensors and big data in basic material extraction. The adoption of Internet of Things technologies could be very high in this industry, perhaps 80 percent to 100 percent," the report states. “Some of the best positioned companies (for the Internet of Things) may be suppliers of big data and analytical software that can help extract meaning from the enormous flows of data.” Advanced oil and gas exploration and recovery. “Businesses that participate in the energy value chain, especially energy companies and oil-field services, could find enormous global opportunities in advanced oil and gas exploration and recovery. These players … should master big data analytics to improve their research. Shell, for example, is already collecting up to a petabyte (1 million gigabytes) of geological data per well using its advanced seismic monitoring sensors (co-developed with Hewlett-Packard), and plans to use the sensors on 10,000 wells.” McKinsey acknowledges that many variables at play (such as energy prices) could affect the economic impact of the dozen disruptive technologies in the report, leading the authors to establish an impact range of $14 trillion to $33 trillion by 2025. Overall, though, the report sounds a decidedly bullish note. “While we agree that significant challenges lie ahead, we also see considerable reason for optimism about the potential for new and emerging technologies to raise productivity and provide widespread benefits across economies,” the executive summary says. But Meta Brown, a business analytics expert, consultant and speaker, says she has “come to believe that no technology is, in itself, disruptive.” “Disruption, dramatic change in the way that work is done, or how everyday life is lived, happens when a widespread need is met by the right mix of technology with economic and social conditions,” Brown says. “Can big data fuel technological disruption? Only if the conditions are right. “I certainly agree that some of these technologies would require data to function effectively, and if the scale is large enough, that means big data,” she says. “However, I am not yet convinced that economic and social conditions will drive massive demand for these technologies. Industrial robots, for example, have been available for some time, yet it is still a minority of factories that use them. The mere existence of new and better robots won't change that, unless there are other changes in conditions as well.” Chui and his McKinsey colleagues also wrote an oft-cited report from 2011 that described big data as "the next frontier for innovation, competition and productivity" and projected a coming skills shortage among data professionals and managers who can make data-driven decisions. Contributing Editor Christopher Nerney (cnerney@nerney.net) is a freelance writer in upstate New York. Follow him on Twitter @ChrisNerney.

The post McKinsey Report: Big Data at Center of Disruptive Technologies appeared first on Data Informed.

Advertisers Push for Data to Measure Audiences in Multiple Channels

Cyril Tuohy — Thu, 23 May 2013 21:32:19 +0000

PHILADELPHIA—In case advertisers hadn’t noticed, the era of the double click is dead. In its place: a world of quadruple clicks and multiple finger swipes, a multiscreen shopping experience in which advertisers need to target their message across a plethora of platforms to make a sale.

Related Stories

How marketers use game mechanics to engage consumers and analyze behavior. Read the story »

Inside Wharton’s Research Center for Customer Analytics. Read the story »

Guide to Sentiment Analysis. Read the story »

Integrating social analytics with CRM uncovers new customer views. Read the story »

For advertisers, the challenge is how to measure the effectiveness of their ad dollar when consumers hear about a product on television, research it two days later on a mobile phone, comparison shop on other websites using a desktop, before finally closing the sale on a tablet on their commute home from work. “A lot of people are seeing ads on mobile phones but when they buy the product they are using the larger screen on PC or laptop,” said Anindya Ghose, professor of Information, Operations and Management Sciences at New York University’s Leonard N. Stern School of Business in New York. Ghose was one of several panelists who spoke at a May 16 conference at the Wharton School’s Future of Advertising Program which explored innovations in measuring advertising effectiveness. For managers looking at the click-through and sales data, new shopping habits anchored in multiscreen behavior pose a quandary. Where does management plunk down the ad dollar to get the most bang for the buck? Mobile ads influence not only mobile purchasing propensities, but Web purchasing habits as well. Mobile clicks precipitate conversion to a sale on the Web, and Web clicks precipitate conversion to a sale on mobile platforms, said Ghose. As a result, there’s an increase in sales when Web and mobile ads are active on both devices within the same time period, and managers who spend their entire ad budgets on the mobile channel only risk missing out on reaching consumers through the PC channel. Ghose, who spoke on the interdependence between Web and mobile advertising, said takeaways from his research were that mobile ads influence cross-media conversions to a sale and that ad campaigns need to incorporate the relationship buyers have across multiple devices. The multichannel model is where the future is headed, said Justin Petty, vice president of global media solutions and partnership with DunnhumbyUSA, the Cincinnati-based branch of the British data and customer analysis company. Exposing consumers to multiple channels before they make a buying decision isn’t new, of course, but new Web-based channels—mobile and social media—require new data models to allow marketers to track consumers who surf from a Facebook ad to a Google search to pressing “buy” at the online shop, said Eva Anderl, a Ph.D. candidate at Universitat Passau in Germany. The trick is for companies to figure out the relative value of each click as consumers jump from one channel to another.

Field Notes from Ad Researchers

The Wharton Future of Advertising conference was filled with young executives eager to share their latest findings using Markov simulations and regression models to explore connections between incremental purchases and the frequency of online display ads, for example. Among the projects researchers discussed were the following:

More online ads do not necessarily lead to more sales. The connection between an incremental online purchase and a higher frequency of ads remains a tenuous proposition, the researchers found, though more ads lead to more searches for products and services.

Online display ads are generally effective, yes, but there are a host of variables – dependent and independent – involved in determining just how effective, said Garrett Johnson, a doctoral candidate in economics at Northwestern University.

Johnson, who conducted research with three million customers in conjunction with Yahoo and a major retailer, said variables determining the effectiveness of an online ad range from proximity of where the ads have run, how recently they have been shown, how loyal the audience is to a product, and the income segment of the buyers.

In all cases, he said, it pays for businesses to invest in a few statistical calculations before assuming that more online display ads will translate into more sales.

Traditional research methods still apply. Businesses need to establish baselines and use “control” ads to benchmark their campaigns, and ad managers who target ads more carefully stand to improve future campaigns, he said, adding that focus groups remain a good use of money. “Test the effectiveness of the ads with focus groups,” Johnson said.

Watch out for measurement errors. Randall Lewis, senior economic research scientist at Google Inc., warned of “attenuation bias,” caused by measurement errors when looking at ad effectiveness. “There are so many areas where attenuation bias can get in,” he said. “Try and pin down specific numbers.”

Search continues to matter—a lot. There is a very strong relationship between ads and search, Lewis said, as people are drawn to a product and take it upon themselves to find out more about a particular product, even if they may not buy it in the end.

In the auto segment, for instance, Google sees an immediate increase in search across all competitor brands when an ad for Acura pops up, he said.

Advertising influence depends on the channel. Format is yet another variable that affects an ad's effectiveness. Porting over a successful Web-only campaign is no slam dunk when it comes to TV or other channels, said David Reiley, a Google research scientist.

TV ads, he said, are much more “in-your-face” than an online ads. “I’d be cautious about extending online ads to other forms.”

Justin Petty of DunnhumbyUSA said that people seeing ads on Facebook—in effect PC users—are price sensitive. TV ads, by contrast, are pushed onto viewers, which is a completely different audience, and companies are asking how to approach these different audiences through their advertising.

“That’s an interesting question for marketers,” Anderl said, particularly small and midsize companies who still use simpler models that rely on standard advertising performance measures which are no longer adequate. Her model, implemented at the German tracking company intelliAd, helps allocate advertising dollars as a function of the value of the exposures, and gives greater weight to search engine optimization, Anderl said. Andy Fisher, senior vice president and chief analytics officer for Merkle Inc., said one of the big trends in the industry today is for advertisers to incorporate social media into the more traditional media mix and attribution models. Results have by and large been uneven, but the movie industry has had the most success with it so far, he said. Other techniques to measure advertising effectiveness include eye-tracking and emotional responses on people’s faces, said Daniel McDuff, a research assistant with MIT Media Lab. “More and more people are watching ads online and TV online, so there's a two-way relationship where we can collect information on their responses to the content,” he said. That is, though, only if they agree as part of an opt-in process to be tracked, and even McDuff conceded that the “self-selection bias” needs to be taken into account. Video platforms are still another data gold mine for retailers and advertisers, said Anuj Kumar, assistant professor of information systems management at the University of Florida’s Warrington College of Business Administration in Gainesville. After the introduction of product videos, which allow shoppers to see what they might look like wearing different pieces of clothing, one midsize women’s apparel retailer saw a 15 percent increase in sales of its featured product, and a 31 percent increase in sales of coordinating products, Kumar said. Mining that data could yield a wealth of information about shopping preferences, leading to more effective advertising, even if companies have not exploited video-based data techniques to their fullest potential, Kumar said. Data collected by video can be assimilated relatively cheaply using back-end systems, he said. Yet for all the charts and graphs and the overwhelming rationality that governs advertising effectiveness measurement today, Orlando Wood, managing director at BrainJuicer, an international market research and analytics firm, said the most effective – and profitable – ads are the ones that connect emotionally with people. When it comes to making a decision, “we think much less than we think we think,” Wood said, gleefully casting aside – if only for a moment – all the day’s statistical analysis. “The way advertising works is by making you feel something for the brand and nudging you in its favor.” Cyril Tuohy is a freelance business writer based near Philadelphia. He can be reached at cyriltuohy@gmail.com. Home page photo of family watching TV via the National Archives and Records Administration.

The post Advertisers Push for Data to Measure Audiences in Multiple Channels appeared first on Data Informed.

Pattern, Open Source Framework, Aims to Accelerate Analytics on Hadoop

Cindy Waxer — Thu, 23 May 2013 19:04:25 +0000

Apache Hadoop may be able to work through huge datasets, but it can also present its fair share obstacles when it comes to deployment. “Hadoop was designed to solve two problems: One, to remove the question of where to store data, and two, where to perform computing. Companies don’t need to ask these questions anymore; they just put it on Hadoop,” said Chris Wensel, CTO and founder of big data application provider Concurrent. “However, the problem is Hadoop is extremely hard to use. Many data scientists aren’t even in production yet with their models.”

Related Stories

Concurrent’s Lingual designed to let SQL developers run big data applications on Hadoop. Read the story »

How predictive model markup language puts big data to work faster for business. Read the story »

Guide to Predictive Analytics. Read the story »

Innovative relational databases create new analytics opportunities for SQL programmers. Read the story »

The roadblock developers face, said Wensel, is that while it’s relatively easy to build a Hadoop cluster, converting it into a data analysis system requires a mix of sophisticated data mining tools such as R, MicroStrategy and SAS and languages like Pig, Hive and HBase from a whole host of vendors including Hortonworks, MapR Technologies and Cloudera. Unfortunately, according to Wensel, many of these tools are meant for “data cleansing and scoring, not for preparing machine-learning applications for production.” As a result, he said, “A lot of code must be written to turn cool models into production-oriented applications that users can trust.” Now Wensel believes he’s discovered a way to ease and expedite Hadoop deployments with Pattern, an open source scoring engine. A scoring engine sorts through and prioritizes a model’s huge volumes of data to make predictions about future behavior – a time-consuming and laborious yet necessary process. Pattern improves this process by allowing users to create a model using a programming language such as R. However, because R doesn’t always run efficiently on Hadoop, users simply use PPML, a standards-based language for exporting, to carry the model from R over to Hadoop. Next, Pattern converts the PPML model to a Cascading application where developers can begin building enterprise-grade big data applications. Cascading is an application framework for Java developers working with Hadoop, also developed by Concurrent. By quickly deploying machine-learning applications on Hadoop, Pattern can significantly reduce development time from weeks or months into “hours, if not minutes,” said Wensel. What’s more, because Pattern runs on Concurrent’s Java application framework Cascading, companies can leverage existing investment in big data technology such as Java, SQL and predictive modeling. That’s good news to Pattern production use cases like WellPoint, a managed health care provider that’s test-driving Pattern to provide analytics used by federal agencies. Social website Airbnb is also using Pattern to better connect travellers to home-dwellers with room to spare. And then there’s early Pattern evaluation user AgilOne, a company that makes cloud-based predictive analytics and business analytics tools. Antony Arokiasamy, senior software architect at AgilOne, said in a statement that using Pattern allow for the deployment of a variety of machine-learning algorithms for its predictive marketing tools. “As a self-service SaaS offering, Pattern allows us to evaluate multiple models and push the clients’ best models into our high performance scoring system. The PMML interface allows our advanced clients to deploy custom models,” Arokiasamy said. Pattern’s release follows closely on the heels of the unveiling of Concurrent’s Lingual, an open source ANSI-standard SQL engine that runs on top of Cascading. Lingual lets data scientists and developers with basic SQL skills build applications on Hadoop without any training in MapReduce. Instead, users are able to run MapReduce applications from standard SQL queries. These programming tools are designed to make creating predictive models easier. But heavy coding and a hodgepodge of sophisticated tools aren’t the only hindrances to Hadoop deployments. Erik Jarlstrom, vice president of technology solutions at Dataguise, a provider of data privacy tools, has long warned of the data security dangers that keep Hadoop applications in pilot. Hadoop deployments, with their wide variety of data classifications including log files, structured data and mixed data types, can be difficult to secure. Because of this, Jarlstrom has issued a wake-up call to those that aren’t taking advantage of the tools needed to secure the “serious holes” plaguing Hadoop deployments. Whether the hesitation springs from data security concerns or a painstakingly long development cycle, there is one thing developers can agree on: “You can have the best model in the world after spending months creating it, but if you can’t put it in production, it’s useless,” said Wensel. Cindy Waxer, a contributing editor who covers workforce analytics and other topics for Data Informed, is a Toronto-based freelance journalist and a contributor to publications including The Economist and MIT Technology Review. She can be reached at cwaxer@sympatico.ca or via Twitter @Cwaxer.

The post Pattern, Open Source Framework, Aims to Accelerate Analytics on Hadoop appeared first on Data Informed.

Enterprises Value Data Pros With Multiple Skills, Dice Survey Shows

Christopher Nerney — Thu, 23 May 2013 00:29:25 +0000

At the MIT Sloan CIO Symposium on May 22, John Roese, the CTO of technology company EMC, was one of several IT executives to cite the need for skilled data pros. "When it comes to analytics, there’s not a large talent pool. The actual tools are new. It’s a challenge to find 1,000 Hadoop experts," Roese said. "They just aren’t available."

Related Stories

Five ways for tech startups to attract analytics talent. Read the story »

Dice launches analytics service for IT job recruiters. Read the story »

Find more articles and resources in Data Informed's Skills section. Read the story »

It’s no secret that professionals with skills related to big data and analytics technologies are in demand now and in the immediate future. And strong demand for talent usually means higher salaries. But getting beyond that basic economic equation, what specific analytics skills or areas of knowledge are commanding the highest salaries? According to technology jobs and career site Dice.com, Hadoop skills are most in demand among enterprises seeking analytics talent, with the average annual salary for an analyst with Hadoop skills averaging $115,062. The more generic “big data” job category averages $113,739 a year, while jobs in the Linux category average $90,853, Dice.com reports. Compare that to the $85,600 average salary across all tech categories (according to the latest Dice Salary Survey), and it’s clear that demand for workers with analytics-related skills is extremely robust. But Dice says hiring managers typically want candidates with Hadoop experience to also have additional skills. “The number one combination by a large margin,” writes Dice.com’s Howard Lee, is “Hadoop and Java—not too surprising given it’s a Java-based framework.” Second among specific skills demanded by hiring professionals of Hadoop pros is NoSQL. The Dice Salary Survey reports that professionals with Hadoop, NoSQL and MongoDB experience all have annual salaries “north of $100,000,” Lee says. Following NoSQL, the skills most in demand among hiring pros seeking candidates with Hadoop are Map Reduce, Pig and Hive. Dice.com gathered the data through Open Web, the company’s new service that compiles social profile information from millions of professionals to create aggregate profiles. Though Dice designed Open Web for enterprises seeking tech talent, the job site also uses it to track skills trends that may be below the surface. "Companies are beginning to understand there is huge value locked away in their data,” says Lee, who is chief architect of Open Web. “Big data professionals that can help companies leverage these assets to better understand their customers are seeing increasing demand for their skills. This trend is set to continue as big data technologies continue their path towards mainstream adoption." Gathering additional salary data for positions requiring Hadoop and another tech skill is difficult at this point, says Dice.com spokesperson Rachel Ceccarelli. “Big data is a newer category and there aren’t a lot of tech professionals with some of these skills, let alone the combo, which is why they’re so in demand,” she says. “We don’t have a large enough population to draw from for salaries yet.” As for which industries are most eager to hire big data professionals, Ceccarelli says it’s a long list. “We know that companies in online retail, financial services, software, and telecommunications are all looking for tech talent with Hadoop and big data experience,” she says. “It’s hard to say which industry is tops, but it’s clear that as firms continue to analyze large quantities of data, big data talent will be in demand.” Nonetheless, one big takeaway from Dice’s data is that of the four horsemen of technological transformation – mobile, cloud, social and data – those professionals with data skills are being paid the most. Average salaries for technologies associated with cloud and virtualization are just under $90,000, while mobile salaries are close to $80,000, Dice says. Another jobs site, Indeed.com, also shows high pay for big data professionals. A search on “Hadoop” shows an average salary of $107,000, while a search on “NoSQL” shows an average salary of $100,000. Here are some average salaries from Indeed.com for jobs requiring Hadoop and another skill: • Hadoop and Java: $103,000 • Hadoop and NoSQL: $107,000 • Hadoop and MongoDB: $118,000 • Hadoop and Map Reduce: $84,000 • Hadoop and Pig: $107,000 • Hadoop and Hive: $109,000 Contributing Editor Christopher Nerney (cnerney@nerney.net) is a freelance writer in upstate New York. Follow him on Twitter @ChrisNerney. Home page photo by Flickr user bgottsab, used under Creative Commons.

The post Enterprises Value Data Pros With Multiple Skills, Dice Survey Shows appeared first on Data Informed.

A Real-Time Mind-Set for Marketing

David Meerman Scott — Tue, 21 May 2013 20:19:27 +0000

When I speak with executives around the world about social, many think of their kids’ Facebook or Twitter and what you had for lunch, deciding that social is frivolous at best and a dangerous time-waster at worst. But social business is not just about social media. In discussing enterprise marketing strategy, I recommend not using the word “social” at all and instead substitute “real-time.” An immensely powerful competitive advantage flows to organizations with people who understand the power of real-time information. What are people doing on your website right now? Has someone just praised you on Facebook? Panned you on Twitter? Published a how-to video about your product on YouTube?

Related Stories

Podcast: The rise of real-time marketing in a data-driven world. Read the story »

Integrating social analytics with CRM uncovers new customer views. Read the story »

Executives understand real-time. And they are eager to implement real-time ideas. Conventional vs. Real-Time The conventional business approach to marketing favors a campaign (note the war metaphor) that requires people to spend weeks or months planning to hit targets. Agencies must be consulted. Messaging strategies must be developed. Advertising space and time must be bought. Conference rooms and refreshments must be prepared for press conferences. Do you serve them sushi or sandwiches? The real-time mind-set recognizes the importance of speed. It is an attitude to business (and to life) that emphasizes moving quickly when the time is right.

Three Tactics for Real-Time Engagement

1. Connect with customers now.

Finding ways to interact with customers on a regular basis and in real time is something of an art form. But if you have the right creative approach social media can now make communication instant, easy and free.

Albion Cafe in Shoreditch, London, sends a tweet (@albionsoven) when baked goods come fresh and hot from the oven. Example: Fresh coconut cake.

Locals subscribe so they know exactly when to pop over.

2. Develop a real-time communications policy.

Develop an effective code of real-time communications and proactively embed it throughout your organization. Guidelines mean employees know they have the freedom to communicate in real time when the opportunity arises. Train it, demonstrate it, discuss it and review it until this becomes second nature to everyone. Have your people internalize it as deeply as the instincts that tell them when it’s safe to turn left at a traffic light (or right if they’re Brits).

IBM’s code is called “Social Computing Guidelines.” The purpose is to provide rules to help employees engage the marketplace and customers in real time—effectively and responsibly.

“A big part of being engaged in the community is feeling comfortable with what you can say and what you can’t say, so we wanted to establish the boundaries," says Tim Washer, who headed social media for IBM when the guidelines were developed. "If you identify yourself as an IBMer, then you need to adhere to the guidelines. The guidance we offer is that, if you give perspective on a topic that has something to do with IBM, we want you to speak as an IBMer.”

3. Enable real-time technology.

To support real-time business, you need technology infrastructure every bit as sophisticated as a financial trading floor. When well-integrated into an appropriate technology backbone these modules work together to feed the dashboard that your marketers, PR professionals, salespeople, and executives use every day.

Developing a real-time mind-set is not an either/or proposition. I'm not saying you should abandon your current business-planning process. Nor do I advocate allowing your team to run off barking at every car that drives by. Focus and collaboration are essential. Large Organizations Need to Work at It The more people you have in an organization, the tougher it is to communicate in real time. In a command-and-control environment where no action can be taken without authority, without consultation, without due process, any individual who shows initiative can expect to be squashed. The challenge is to develop a new balance that empowers employee initiative but offers real-time guidance when it’s needed—like a hotline to higher authority. In a real-time corporate culture everyone is recognized as a responsible adult. If you're the leader, and you want to cultivate a real-time mind-set throughout your organization, tear down the command-and-control mentality. Recognize your employees as responsible adults. Empower them to take initiative. Real-Time Leadership If you’re the leader, and you want to cultivate a real-time mindset throughout your organization, tear down the command-and-control mentality. Recognize your employees as responsible adults. Empower them to take initiative. Give them opportunities to hone their communication skills, give them clear guidelines as to what’s appropriate and what’s not. Scale and media buying power are no longer a decisive advantage. What counts today is speed and agility. While your competitors scramble to adjust, you can seize the initiative, open new channels, and grow your brand. David Meerman Scott (Twitter: @dmscott) is a marketing strategist and the author of several books including Real-Time Marketing & PR: How to Instantly Engage Your Market, Connect with Customers, and Create Products that Grow Your Business Now. He is scheduled to speak at Data Informed’s Marketing Analytics & Customer Engagement event, June 24-25 in Philadelphia.

The post A Real-Time Mind-Set for Marketing appeared first on Data Informed.

L’Oréal USA Signs Outsourcing Deal with IBM for Procurement Analytics

Michael Goldberg — Tue, 21 May 2013 18:45:54 +0000

Mascara was big for L’Oréal USA in 2012. Mega Plus, the latest addition to the Volum’Express mascara line, sold several million units. Combined with new eye shadow and lipstick product launches in L’Oréal’s biggest market, Mega Plus made the French cosmetics giant look good, able to post a 12 percent rise in operating profits on sales of 22.5 billion Euros in a fragile economy. Rich Ullrich also works to make L’Oréal look good, but as vice president of indirect procurement at New York-based L’Oréal USA, he’s concerned with everything that does not relate to the development and manufacture of beauty products—purchases that amount to about 60 percent of the company’s overall spending. And while sales representatives at makeup counters explained how the Mega Plus gel-mousse formulation leads to supple eye lashes, Ullrich was working to embark on a three-year procurement analytics deal with IBM. The initiative’s goal: implement data management and analytics tools, benchmarking studies and improved management practices to cut costs by as much as 12 percent of baseline spending on indirect procurement. Office space. Travel expenses. Consulting services. Marketing and advertising purchases. IT equipment. And more.

Related Stories

Guide to Procurement Analytics. Read the story »

Visualizations highlight data for improving manufacturer’s procurement process. Read the story »

Integrating sustainability into corporate decision-making. Read the story »

More on supply chain analytics. Read more »

L’Oréal is not alone in having a procurement challenge. Enterprises can find it difficult to track purchases, for example, when a company’s size or structure changes with a merger or acquisition, when purchasing is distributed in an organization, or when different vendors use different labels for the same category of office supplies or services. Applying analytics to expenditures is a long-standing use case. Through growth and acquisitions, L’Oréal USA has seen its efforts to control procurement practices fragment, Ullrich said. With multiple versions of enterprise applications in place, getting a unified view of nine different sets of procurement data will allow the company to understand spending behavior, prices paid for goods and services and “then apply levers” to get better results. “This project is really about resource optimization and spending leverage,” Ullrich said. “The three things [required] are the right resources, leveraging your spend and utilizing industry-best tools. Those things make for a successful project.” L’Oréal USA and IBM did not disclose terms of their deal which went into effect earlier this year, except to point out that the amount IBM earns in fees will depend on how much money L’Oréal saves. Other issues Ullrich discussed: • The project will use Emptoris tools to collect, clean, manage and analyze procurement data. An acquisition announced in December 2011, IBM completed its purchase of Emptoris in 2012. • The outsourcing deal makes sense to L’Oréal USA because it means the cosmetics company avoids having to spend the time to build, recruit and train its own analytics team. At the same time, Ullrich said, L’Oréal can choose to bring the project in-house after three years if it chooses based on the terms of the licensing deal with IBM. • Benchmarking is a key element of the project. The ability to compare L’Oréal’s procurement data and history to other enterprises is a priority. Such benchmarking allows L’Oréal to examine the pricing for commodities and to assess how L’Oréal can get better deals, he said. The deal with IBM allows L’Oréal to analyze a broad view of many companies’ procurement experiences, Ullrich said. • Eventually, L’Oréal USA plans to connect its procurement analytics with its SAP ERP system. The issue of strategic sourcing will be a major focus of the L’Oréal USA project, said Bill Schaefer, IBM’s vice president, supply chain and procurement services. “Strategic sourcing is the front end of the process. It’s about identifying the suppliers who I want to do business with, the bidding and negotiating, and going to contract with them.” Compliance is another element of the project, and another classic issue in procurement, he said. Companies set policies, such as traveling with a preferred airline for pre-negotiated rates, but fail to follow through. Many companies see less than 50 percent compliance, “so they leave tons of money on the table,” Schaefer said. “Part of the scope of this project is to help [L’Oréal USA] drive much higher compliance.” He said the project will deploy the Compliance Analytics Tool developed by IBM Research that mines data on millions of day-to-day transactions. “When someone places a purchase order or wants to buy something, [the tool] looks for problem areas, if you order from the wrong supplier or you are not getting the right prices. And it helps pinpoint where the problem areas are and then goes back and addresses it.” Maybe the problem is that employees were not aware of the company’s purchasing policies. Or maybe the contract with a vendor needs to change. “The idea behind it is to make sure L’Oréal sees the benefit of all those contracts,” he said. “Many people might not appreciate the compliance issue, but it’s a huge issue,” said Schaefer, who was a procurement executive at IBM before leading its service offering. “A typical company may be doing millions of transactions. People buying office supplies, and all kinds of things. How are you going to find problems? If you run a standard report, who did I place order with? But you’d have to manually go through that, and then spot suppliers not on your list.” Michael Goldberg is editor of Data Informed. Email him at michael.goldberg@wispubs.com. Follow him on Twitter at @MGoldbergatDI.

The post L’Oréal USA Signs Outsourcing Deal with IBM for Procurement Analytics appeared first on Data Informed.

Program Seeks to Merge Electronic Health Data and Genomic Data

Wendy Meyeroff — Mon, 20 May 2013 15:34:08 +0000

There’s an important effort underway among health care data experts to enable clinicians and medical researchers to share the same data for analytics to improve patient outcomes. At issue is the structure of electronic health records (EHR) that were originally designed to be used in day-to-day patient care and are not set up to handle much bulkier data types such as X-ray images and genomic tests.

Related Stories

Data quality study shows the potential of electronic health records. Read the story »

Health care analytics that get results involve doctors in developing data tools. Read the story »

A prescription for better analytics at Atrius Health. Read the story »

How HR Managers Can Prevent Data Hoarding and Improve Productivity

Cindy Waxer — Thu, 16 May 2013 22:17:51 +0000

The world is producing data at an unprecedented rate. IDC predicts that from 2005 to 2020, the digital universe will grow by a factor of 300, from 130 exabytes to 40,000 exabytes, or 40 trillion gigabytes. While all this data can help companies better target customers, drive sales and identify top talent, too much information can degrade the effectiveness of data-driven processes and impede employee productivity.

Related Stories

4 tips for evaluating workforce analytics software. Read the story »

Guide to Human Resources Analytics. Read the story »

HR executives: Analytics role needs higher profile. Read the story »

More on human resources. Read more »

Fortunately, there are steps HR leaders can take to control the data deluge facing so many companies. Just ask Amber Simonsen. Simonsen has spent the better part of her career as a consultant and senior project manager convincing companies of the “legal risks and risks to co-workers” of being a data pack rat. She offers these tips for making data hoarding a thing of the past. 1. Whittle down your data. According to Simonsen, too much data can actually stand in the way of in-depth analytics. “You get to a point where you have to ask yourself what you plan to produce with the data,” says Simonsen. “We’re all so data-hungry that sometimes our eyes are bigger than our stomach. Organizations have a ton of data – there’s no end in sight – but is data from five years ago really going to provide you with anything valuable?” Simonsen suggests that deleting age-old data (10 years old or more) from an analytics system can help algorithms perform better and produce more accurate results that haven’t been biased by a system’s historical data. 2. Establish and enforce data policies. Meaningless data stored on workers’ personal drives is the bane of any HR professional’s existence, especially when the goal is to purge unnecessary information. Incomplete, preliminary or draft documents that are stored in employees’ email inboxes or document folders can add up over time, “preventing employees from sharing important digital resources and causing co-workers to lose track of versions of documents which can be really frustrating,” warns Simonsen. More than simply frustrating, data hoarding can also impact employee productivity. “If you can’t find the latest version of a document because it’s living on someone’s personal drive, you’re going to wind up working on an older version. That’s such a waste of time, all because someone wants to maintain control or doesn’t trust the system to put a file on SharePoint.” Another data hoarding landmine is legal liability. “You don’t know what’s sitting on personal drives and you don’t know what’s in an email yet you’re held responsible for it as an organization,” warns Simonsen. For example, a company may hand over all the information it believes it has in response to an e-discovery request, not realizing there’s more to be had on an employee’s personal drive. So how can HR help reduce duplicate documents and minimize the risk of legal liabilities? Encouraging the use of collaborative software tools such as Microsoft SharePoint or custom wikis can prevent employees from storing dozens of document versions on their personal computers. What’s more, a BYOD (bring your own device) policy can ensure workers only store sensitive data on corporate-issued smartphones and other mobile devices. 3. Lead efforts to educate employees about data hoarding risks. Traditionally, HR has taken a backseat to IT when it comes to educating employees on the perils of data hoarding. But that’s a huge mistake, according to Simonsen. “HR can play a critical role in helping people understand what data truly is valuable, the cost to maintain documents and the risks to the organization,” she says. “When an IT department is involved, it’s a much less warm-and-fuzzy process. Their approach is to say, ‘We’re going to take your files and we’re going to put them here.’ For some employees, it feels as if they’re being violated.” HR, however, can take a much more people-friendly than data-centric approach by educating employees on the dangers and costs of storing too much data, as well as establishing benchmarks on how long workers should archive certain types of data. Other Rules of Thumb So how much is too much data? According to survey results from the Compliance, Governance and Oversight Council, a consortium of 2,300 legal and IT professions, organizations on average need to archive about 2 to 3 percent of their data for legal responsibilities, 5 to 10 percent to meet regulatory requirements, and 25 percent for business analysis and insights. While it’s safe to say that most companies overshoot these standards, some experts argue that there is a time and a place for data hoarding. In a blog posting entitled, “The Case for Data Hoarding,” Phil Simon, author of Too Big to Ignore: The Business Case for Big Data, argues that data hoarding makes sense under two conditions: 1. If an organization is using data tools like Hadoop and NoSQL; and 2. If a company actually does something valuable with its data. Simon even goes so far as to suggest that yesteryear’s arguments against data hoarding, such as the high cost of storage, no longer apply in this era of new storage technologies. Nevertheless, for companies that fail to purge their petabytes, even every now and then, they’re likely to discover that you can have too much of a good thing. Cindy Waxer, a contributing editor who covers workforce analytics and other topics for Data Informed, is a Toronto-based freelance journalist and a contributor to publications including The Economist and MIT Technology Review. She can be reached at cwaxer@sympatico.ca or via Twitter @Cwaxer. Home page photo of papers by Flickr user grace_kat, used under a Creative Commons license.

The post How HR Managers Can Prevent Data Hoarding and Improve Productivity appeared first on Data Informed.

Manufacturers See Savings, Service Boost in Use of Transportation Data

Ian B. Murphy — Thu, 16 May 2013 21:30:55 +0000

ORLANDO – For companies that manufacture and move goods, transportation is an area where improvement can mean major cost savings and better service for their customers. At SAP’s SAPPHIRE conference, two customers of the German business software giant described their approaches for improving their use of data to address the transportation problems plaguing their organizations. Woodgrain Millwork, a Utah-based lumber supplier and custom moulding manufacturer, is gearing up for a June launch of SAP’s Transportation Management software. The company is starting by centralizing transportation for seven of its 17 sites that were previously siloed and working independently.

Related Stories

Partners in SAP ecosystem prepare for real-time supply chain analytics. Read the story »

Effective manufacturing analytics requires cross-functional data sharing. Read the story »

Guide to Procurement Analytics. Read the story »

More on supply chain analytics. Read more »

These seven mills ship lumber to retailers like Home Depot and independent lumber yards. Connie Moylan, the CIO of Woodgrain Millwork, said by bringing the operations together and managing transportation with a single piece of software they expect to save 15 percent on transportation costs. “We did our evaluation for where the best ROI was for a new system, and it was by far in transportation management,” Moylan said. With the new software, and by outfitting each truck with GPS systems, Moylan’s company will be able to track the companies 3,500 trucks and drivers. “It allows us to see and manage where drivers are going,” she said. “It also allows us to monitor driver behavior, whether they’re braking too hard or driving too fast. That’s where some of the cost savings lies, on equipment wear and tear.” The other savings come from fuel consumption, where by better understanding routes and customer demand, the company can send fewer trucks. “If you can get stuff moved on two trailers instead of three, then you’ll be a lot more efficient with the amount of fuel you use,” said Bill King, SAP’s solution manager for the Transportation Management software. Moylan said the new system allows for a much more granular view of the data, and when the system expands to all 17 of her company’s shipping sites she expects there will be even more process improvements that become apparent after data analytics. Better Data, Better Results Global manufacturing company Honeywell had a different problem in its North American performance materials and technology (PMT) division. Honeywell PMT manufactures chemicals, plastics and resins and ships about 60,000 deliveries a year to its customers in North America. Tom Ventrudo, a lead systems business analyst at Honeywell, said the company wanted to improve its customer service by being able to create a more accurate delivery schedule. “We did this first and foremost to delight the customer,” Ventrudo said. “But not only are we getting the product to them on time, we’re getting the paperwork on time as well.” Ventrudo said the company found that its previous system, which broke down each continental U.S. state and Canadian province into a single zone, simply wasn’t accurate enough. Deliveries promised in large states or Canadian territories weren’t taking into account true distance; for example Houston and El Paso Texas are 745 miles apart, more than a single day’s drive. So Ventrudo and his team built a system that takes advantage of the Google Map API. By connecting the company’s SAP ERP Central Component software to this data source, the company could get the true distance to within a hundredth of a mile. With this as a foundation, the company built additional layers to handle exceptions and to help customer service representatives give accurate delivery estimations. The result of adding Google’s GIS data to its SAP system resulted in a 53 percent decrease in late deliveries. One of the best parts, Ventrudo said, is that Google’s API is free to anyone who connects fewer than 25,000 times a day. Honeywell’s PMT division only connects about 200 times daily. Email Staff Writer Ian B. Murphy at ian.murphy@wispubs.com. Follow him on Twitter @IBMurphyatDI.

The post Manufacturers See Savings, Service Boost in Use of Transportation Data appeared first on Data Informed.

Partners in SAP Ecosystem Prepare for Real-Time Supply Chain Analytics

Ian B. Murphy — Thu, 16 May 2013 15:00:45 +0000

ORLANDO – Bill McDermott, the Co-CEO of SAP, made it clear at his company’s SAPPHIRE conference that the in-memory database HANA was SAP’s future. “Not only does HANA represent the intellectual renewal for SAP, it’s the platform that will run every single SAP product going forward,” McDermott said at the conference May 14.

Related Stories

Guide to Procurement Analytics. Read the story »

HANA for ERP brings new ROI calculations to the enterprise. Read the story »

SAP pitches HANA platform to analytics startups. Read the story »

SAP, ScaleOut Software, bring in-memory connectors to Hadoop. Read more »

The promise of HANA is speed. The company says its in-memory processing system brings as much as a 1000x improvement to SAP’s business software. While the full vision of a HANA-supported system of real-time views into all the partners in a supply chain might be two or three years away, the potential has players in SAP’s ecosystem laying the groundwork. For technology partners or companies recently acquired by SAP, like the procurement and spend management company Ariba, which means looking at their capabilities differently and thinking about how HANA changes their business model. SAP’s $4.3 billion purchase of Ariba was finalized in October 2012. Ariba currently connects buyers and sellers in its procurement network cloud service that features more than 1 million companies. The company creates efficiencies by creating a single paperless platform for the transactions and provides buyers and suppliers a vast marketplace. “Networked companies are just more successful,” said Greg Spray, Ariba’s vice president of solutions management. “More than 80 percent of the world’s commerce is done with inefficient paper, email and phone calls. All of that can go away. “Our customers can not only see what they’re spending, with whom and how much, but they’re able to see their operational as well, to understand where their bottlenecks are, where things are being held up in the process, what’s making the process less efficient,” Spray said. “All of that data goes into making easier, more efficient collaborations across the buying and selling.” Now as a part of SAP, Ariba can combine its network of buyers and sellers with the German software giant’s supply chain and ERP software. The result, according to SAP’s Hans Thalbauer, the senior vice president of supply chain management, will be a network where one company can run real time analytics on its entire supply chain, from “its supplier’s supplier to its customer’s customer.” The key to that, Thalbauer said, is HANA. “In order to realize the real time supply chain it’s necessary to bring everything into the in-memory environment,” he said. Richard Jabbour, Tata Consultancy Services’ director of SAP enterprise performance management practice, helps companies that have chosen to use SAP software implement their new systems. On May 15 at SAPPHIRE, Jabbour talked about helping a client, document printing and scanning company Lexmark, to implement a real-time supply chain system based on SAP HANA. With the new system, Lexmark can see in real time where all of its products were around the globe and better manage distribution. Jabbour said that a system that Thalbauer described where a company could track not just its own inventory but also view what was happening with suppliers and customers in real time would benefit businesses profitability. “If I can look at the value stream of my supply chain, and I can start to see bottlenecks that maybe my team can help people solve. Then I can get my products in the door quicker than they’re getting in the door today,” he said. “I can push my days in inventory number down.” SAP likely won’t be able to roll out such a system for two or three years, Jabbour said, as HANA still “needs to mature.” SAP reports there are 1,500 HANA users currently, but Jabbour said most companies are still just beginning to think about what’s possible with real time analytical systems in enterprise management systems like supply chain or spend management. Email Staff Writer Ian B. Murphy at ian.murphy@wispubs.com. Follow him on Twitter @IBMurphyatDI. Home page photo from SAPPHIRE exposition floor by Ian B. Murphy.

The post Partners in SAP Ecosystem Prepare for Real-Time Supply Chain Analytics appeared first on Data Informed.