Lisa Morgan's Official Site

Strategic Insights and Clickworthy Content Development

Page 2 of 9

4 Ways to Improve Data Storytelling

Analytical results are often interpreted differently by different people. Sometimes the conclusions presented don’t align with intuition. Differences in experience and expertise can also come into play. An effective way to align thinking is through data storytelling, although there are better and worse ways to do it.

Data storytelling typically includes text, visualizations and sometimes tables to illustrate a developing trend or issue that requires attention if not action. Data storytelling can make the results more memorable and impactful for those who hear it, assuming the presentation is done effectively. Following are a few things to consider

1. Consider the Audience

Data scientists are often considered poor data storytellers because they struggle to align a story with the needs and knowledge level of the audience. Sometimes others are brought in to translate all the technical jargon into something that that is meaningful to business leaders.

Similarly, different parts of a business may require a slightly different focus that uses different language and maybe even different types of data visualizations to have the desired effect, which is understanding analytical results in context.

2. Tell A Story

Effective stories have a beginning, a middle, and an end. The beginning of a story provides context, setting the stage for the story itself. The middle tells the story, and the end usually includes a set of possibilities. Getting the end right is important because insights without action have little value. Are there actionable insights from the data? How can the results be used to drive strategy? In a business context, is there a significant revenue opportunity or an opportunity for cost savings? How much more likely is it that one course of action will succeed versus another? If you provide curated data points and visualizations that support the key points, you can often pre-emptively address the most likely questions and objections.

Effective storytelling also address issues beyond the “what.” Take a sales situation for example. Heads of sales are constantly monitoring progress against sales targets. Let’s say sales fell short or exceeded expectations last quarter. That leads to other questions such as why were sales better or worse than we expected? How could we use those insights to turn the situation around or increase sales even further? How well do we understand our customer base and their requirements? What levers work well and which don’t?

With some solid analytics and effective data storytelling, everyone in the room — the head of sales along with the C-suite or her team can have a common understanding of what impacted the sales results, why, how things are changing and what that means going forward, for the sales team, products, marketing, etc.

Data storytelling should also explain why the analysis was performed, how the analysis was performed, whether hypotheses were proven or disproven in addition to the important findings and what those findings mean for the audience. Some people make the mistake of showing the many steps required for an analysis to demonstrate how challenging the exercise was, which adds little, if any, value.

3. Quality Matters

Great stories can be derailed by simple mistakes, such as misspellings, a lack of focus and a propensity to demonstrate the mastery of a software program to the point of distracting the audience.

Misspellings and grammatical errors tend to be addressed by modern software; however, they don’t always catch everything. Some of them have default settings that limit the amount of text that can be included; however, that’s usually configurable. Sadly, it’ possible to overload stories with so much noise that the audience has trouble staying focused. The point is not clear, in other words. Similarly, trying to get too creative with the colors used in data visualizations can detract the audience’s attention away from the point.

Also consider the presentation of the data in relation to the data itself. On a scale of one to two, a move from one to two reflects a 100% increase. On an actual scale of 25, 50, 100, or 1000, a single-digit increase would appear differently.

4. Be Prepared to Address Alternatives

One of the reasons businesses have placed greater emphasis on analytics versus traditional reporting is the ability to interact with data versus passively consuming it. There is a parallel with data storytelling which is a move away from the traditional and static business presentation format that tends to reserve questions for the end to interactive storytelling in which questions or alternate points of view can be explored live.

Generally speaking, data storytellers should be prepared for questions and challenges, regardless. Why wasn’t something else explored? If a particular variable were added or subtracted, what would the effect be? Of the X possibilities, which is the most likely to see and why?

How SaaS Strategies Are Evolving

Enterprises are subscribing to more SaaS services than ever, with considerable procurement happening at the departmental level. Specialized SaaS providers target problems that those departments want solved quickly. Because SaaS software tends to be easy to set up and use, there appears to be no need for IT’s involvement, until something goes wrong.

According to the Harvey Nash /KPMG 2017 CIO Survey, 91% of the nearly 4,500 CIO and IT leaders who responded expect to make moderate or significant SaaS investments, up from 82% in 2016. The report also states that 40% of SaaS product procurement now happens outside IT.

“IT needs a new operating model,” said Gianna D’Angelo, principal of KPMG CIO Advisory. “CIOs must respond by continuing to focus on operational excellence while adopting a new operating model for IT to drive innovation and value in these changing times.”

Some IT shops are reacting to shadow IT like they reacted to “bring your own device” (BYOD), meaning if you can’t stop it, you have to enable it with governance in mind. However, issues remain.

“In the last three years, we’ve put policies and some governance in place, but it doesn’t matter. You pull out your credit card, you buy an open source application and I have a virus on my network,” said Todd Reynolds, CTO of WEX Health, which provides a platform for benefit management and healthcare-related financial management. “I don’t even know about it until there’s an issue.”

How SaaS pricing is changing

KPMG’s D’Angelo said most SaaS pricing is based on users or by revenue, and that the contract timeframe is three to five years. There has been some movement to shorter timeframes as low as two years.

Sanjay Srivastava, chief digital officer of Genpact, a global professional services company, said his firm sees a shift from user-based pricing to usage-based pricing, which in Genpact’s case takes the form of a per-item charge for a document or balance sheet, for example.

Regardless of what the SaaS pricing model is, SaaS providers are facing downward pricing pressure. According to Gartner, “Vendors are becoming more creative with their SaaS business models to reflect a need to stand out in the fast-growing subscription economy.”

For its part, WEX Health is responding with new services that drive additional revenue. It has also put some usage-based pricing in place for customers that require elastic compute capabilities. “Mobile is killing us,” said Wex Health’s Reynolds. “You’ve given somebody an application to use on their phone 24/7, so they’re starting to leverage that usage so much more. It’s good people are using [our software] more often, but it requires us to have more storage.”

Longer-term thinking is wise

When departments purchase SaaS software, they usually are seeking relief from some sort of business problem, such as multichannel marketing attribution – studying the set of actions that users take in various environments. What business people often miss is the longer-term requirement to share data across disparate systems.

“If you have half on-premises and half in different clouds, you might have a private cloud, some in Azure and some in Amazon because the technology stack is beneficial to the apps,” said WEX Health’s Reynolds. “Pulling all of that together and making it safe and accessible is the biggest challenge from an operational perspective on the IT side.”

While SaaS systems tend to have APIs that help with data exchange, most enterprises have hybrid environments that include legacy systems, some of which do not have APIs. In the older systems, the data dictionaries may not be up-to-date and Master Data Management (MDM) may not have been maintained. So enterprises often face substantial data quality issues that negatively impact the value they’re getting from their investments.

“If you really want to get value out of [SaaS] — if you want Salesforce to run CRM and you want it to run sales, integrated, and it still has to be connected to ERP — each thing has to be connected,” said Genpact’s  Srivastava. “There’s a lot of back and forth. Planning for that back and forth, and planning well, is really critical.”

Part of that back-and-forth is ensuring that the right governance, compliance and security controls are in place.

Bottom line

There’s more to SaaS investments than may be obvious to the people procuring them. At the same time, IT departments can no longer be the sole gatekeepers of all things tech.

“The challenge for CIOs is enormous, the stakes are large and change efforts of this magnitude take years, but transforming the IT operating model can be done,” said KPMG’s D’Angelo. “Complicating the effort is that IT must continue to support the existing portfolios, including retained infrastructure and legacy applications, during the transformation.”

This means that, for a period of time, IT will have to use a hybrid model comprising both the project-oriented, plan-build-run approach and the next-generation, broker-integrate-orchestrate approach, D’Angelo added.

Tips for Ensuring Winning SaaS Strategies

SaaS software is not a one-size-fits-all proposition. Costs and benefits vary greatly, as do the short-term and long-term trade-offs. Following are a few things you can do along the way to ease the transition.

If you’re just starting out, chances are that most if not all of the software you procure will be SaaS because that’s the way things are going. In addition, SaaS allows for an economic shift to relatively low-cost subscriptions that include upgrades and maintenance (an operational expenditure). This is instead of substantial up-front, on-premises software investments that require subsequent maintenance investments and IT’s help (a capital expenditure). Regardless of what type of software you choose, though, it’s wise to think beyond today’s requirements so you have a better chance of avoiding unforeseen challenges and costs in the future.

If you’re piloting a new type of software, SaaS is probably the way to go because you can usually experiment without a long-term commitment. However, be mindful of the potential integration, security and governance challenges you may encounter as you attempt to connect different data sources.

If you’re in production, you’ll want to continuously assess your requirements in terms of software models, integration, compliance, governance and security. As you continue your move into the cloud, understand what’s holding you back. Finance and HR, for instance, may still hesitate to store their sensitive data anywhere but on-premises. For the foreseeable future, you’ll probably have a hybrid strategy that becomes more cloud-based with time.

At each stage, it’s wise to understand the potential risks and rewards beyond what’s obvious today.

Why Surveys Should Be Structured Differently

keyboard-417093_1280If you’re anything like me, you’re often asked to participate in surveys.  Some of them are short and simple.  Others are very long, very complicated, or both.

You may also design and implement surveys from time to time like I do.   If you want some insight into the effectiveness of your survey designs and their outcomes, pay attention to the responses you get.

Notice the Drop-off Points

Complicated surveys that take 15 or 20 minutes to complete tend to reflect drop off points at which the respondents decided that the time investment required wasn’t worth whatever incentive was offered.  After all, not everyone actually cares about survey results or a  1-in-1,000 chance of winning the latest iPad, for example.  If there’s no incentive whatsoever, long and complicated surveys may  be even less successful, even if you’re pinging your own  database.

A magazine publisher recently ran such a survey, and boy, was it hairy.  It started out like similar surveys, asking questions about the respondent’s title, affiliation, company revenue and size.  It also asked about purchasing habits – who approves, who specifies, who recommends, etc. for different kinds of technologies.  Then, what the respondent’s content preferences are for learning about tech (several drop-down menus), using tech (several drop-down menus), purchasing tech (several drop-down menus), and I can’t remember what else.  At that point, one was about 6% done with the survey.  So much for “10 – 15 minutes.”  It took about 10 or 15 minutes just to wade through the first single-digit percent of it.  One would really want a slim chance of winning the incentive to complete that survey.

In short, the quest to learn everything about everything in one very long and complex survey may end in more knowledge about who took the survey than how how people feel about important issues.

On the flip side are very simple surveys that take a minute or two to answer.  Those types of surveys tend to focus on whether a customer is satisfied or dissatisfied with customer service, rather than delving into the details of opinions about several complicated matters.

Survey design is really important.  Complex fishing expeditions can and often do reflect a lack of focus on the survey designer’s part.

Complex Surveys May Skew Results

Overly complicated surveys may also yield spurious results.  For example, let’s say 500 people agree to take a survey we just launched that happens to be very long and very complex.  Not all of the respondents will get past the who-are-you questions because those too are complicated.  Then, as the survey goes on, more people drop, then more.

The result is that  X% of of the survey responses at the end of the survey are not the same as X% earlier in the survey.  What I mean by that is 500 people started, maybe 400 get past the qualification portion, and the numbers continue to fall as yet more complicated questions arise but  the “progress bar” shows little forward movement.  By the end of the survey, far less than 500 have participated, maybe 200  or 100.

Of course, no one outside the survey team knows this, including the people in the company who are presented with the survey results.  They only know that 500 people participated in the survey and X% said this or that.

However, had all 500 people answered all the questions, the results of some of the questions would likely look slightly or considerably different, which may be very important.

Let’s say 150 people completed our  survey and the last question asked whether they planned to purchase an iPhone 7 within the next three months.  40% of them or 60 respondents said yes.  If all 500 survey respondents answered that same question, I can almost guarantee you the answer would not be 40% .  It might be close to 40% or it might not be even close to 40%.

So, if you genuinely care about divining some sort of “truth” from surveys, you need to be mindful about how to define and structure the survey and that the data you see may not be telling you the entire story, or even an accurate story.

The point about accuracy is very important and one that people without some kind of statistical background likely haven’t even considered because they’re viewing all aggregate numbers as having equal weight and equal accuracy.

I, for one, think that survey “best practices” are going to evolve in the coming years with the help of data science.  While the average business person knows little about data science now, in the future it will likely seem cavalier not to consider the quality of the data you’re getting and what you can do to improve the quality of that data.  Your credibility and perhaps your job may depend on it.

In the meantime, try not to shift the burden of thinking entirely to your survey audience because it won’t do either of you much good.  Think about what you want to achieve, structure your questions in a way that gives you insight into your audience and their motivations (avoid leading questions!), and be mindful that not all aggregate answers are equally accurate or representative, even within the same survey.

What Retailers Know About You

Retailers now have access to more information than ever. They’re using loyalty cards, cameras, POS transaction data, GPS data, and third-party data in an effort to get shoppers to visit more often and buy more. The focus is to provide better shopping experiences on a more personalized level. Operationally speaking, they’re trying to reduce waste, optimize inventory selection, and improve merchandising.

The barrier to personalized experiences is PII (personally identifiable information), of course.

“[What retailers know about you] is still largely transaction-based,” said Dave Harvey, VP of Thought Leadership, branding and retail services provider Daymon. “Information about lifestyles, behaviors and attitudes are hard for retailers to get themselves so they partner with companies and providers that have those kinds of panels, especially getting information about how people are reacting through social media and what they’re buying online.”

Transactions Are Driving Insights

The most powerful asset is a retailer’s transactional database. How they segment data is critical, whether it’s transaction-based reach, frequency, lifestyle behaviors, or product groupings. Retailers can identify how you live your life based on the products you buy.

“The biggest Achille’s heel of transaction data, no matter how much you’re segmenting it and how much you’re mining it, is you’re not seeing what your competitors are doing,” said Harvey. “Looking at transaction data across your competition becomes critical.”

As consumers we see the results of that in offers, which may show up in an app, email, flyer or coupons generated at the POS.

Social media scraping has also become popular, not only to gauge consumer sentiment about brands and products, but also to provide additional lifestyle insight.

Some retailers are using predictive and prescriptive analytics to optimize pricing, promotions and inventory. They also have a lot of information about where their customers are coming from, based on credit card transactions. In addition, they’re using third party data to understand customer demographics, including the median incomes of the zip codes in which customers live.

They’re Watching Your Buying Patterns

Retailers monitor what shoppers buy over time, including items they tend to buy together, such as shirts and ties or eggs and orange juice. The information helps them organize shelves, aisles, and end caps.

“There’s a lot of implications for meal solutions and category adjacencies, how people are shopping in the store, how that might lead a retailer test way to offer a right combination of products to create a solution somewhere in the store,” said Harvey. “You can’t be everything to everyone, so how can the information help you prioritize where to focus? The information you can mine from your transaction database, your loyalty card database can help you become more efficient.”

Information about buying patterns and price elasticity allows retailers to micro-target so effectively that shoppers visit the store more often and spend more money.

They May Know How You Shop the Store

Shopping carts and baskets are a necessary convenience for customers, although the latest ones include sensors that track customers’ paths as they navigate through the store.

“They can get the path data and purchase data about how much time you spend at stations, and they can use it to redesign the store or and get you move through the store much more because they know the more you move through the store the more you buy,” said PK Kannan, a professor of marketing science at the University of Maryland’s Robert H. Smith School of Business.

Retailers also use cameras to optimize merchandising to better understand customer behavior including where they go and how long they stay. Now they’re also analyzing facial expressions to determine one’s state of mind.

Driving Business Value from Analytics

Different kinds of analytics result in different ROI. If a retailer is just starting out, Kannaan recommends starting with loyalty cards since other types of data capture and analysis can be prohibitively expensive and the analysis can be cumbersome.

“The ROI on loyalty cards is pretty good,” said Kannaan. “The initial ROI is going to be high and then as you go into more of these cart or visual data, video data, your ROI is going to level off.”

Strategies also differ among types of retailers. For example, a specialty retailer will want data that provides deep insight into the category and shoppers of that category versus a store such as Walmart that carries items in many different categories.

“If you’re a retailer trying to sell a ton of categories you want to understand how people are talking about their shopping experience,” said Harvey. “There’s still a lot of untapped opportunity in understanding social media as it relates to doing better analysis with retailers.”

They’re Innovating

Retailers are working hard to understand their customers, so they can provide better shopping experiences. While personalization techniques are getting more sophisticated, there’s only so far they can go legally in many jurisdictions.

Kannan said a way of getting around this is to take all the informational content, remove any PII, and then extract the resulting information out of the data.

“It’s like I’m taking the kernel from this thing because I don’t have the space to store it and keeping it is not a good policy, so I am going to keep some of the sufficient statistics with me and as new data comes in, I’m going to combine the old data with new data and use it for targeting purposes,” said Kannan. That’s becoming more of a possibility now, and also it’s a reality because data volumes are increasing like crazy. That way I don’t have to store all the data in a data lake.”

AI Has a Foothold in Business, Now for the Next Steps

AI is seeping into different industries, slowly remolding the global competitive landscape. However, most business leaders still don’t know how machine intelligence will impact their businesses.

EY recently published a brief, which focuses the current state of AI. We interviewed Nigel Duffy, EY Global Innovation AI leader who co-authored the document with Chris Mazzei, EY Global Innovation Technologies Leader and Global Chief Analytics Officer.

The brief frames the current state of AI well: “Most organizations aren’t exploiting the potential of AI; they are just at the beginnings of their AI journeys. What should be holding companies back is a lack of talent, but it’s actually a lack of understanding of what’s possible – particularly at the top of large enterprises.”

Addressing the C-Suite disconnect

It’s often hard to imagine the impact new technologies will have on a business. Granted, AI is not new; however, due to recent research and developments, it’s finally at a point where more organizations are either using it in production or experimenting with it.

Some say AI is at an inflection point, namely, at the beginning stages of exponential “hockey stick” growth. If that’s true, the latecomers may find themselves blind-sided by competitors, simply because they didn’t think about and learn first-hand how AI would affect their own companies.

According to Duffy, part of the confusion stems from the fact that AI is a broad set of technologies as opposed to a single, coherent capability. Given the complexity of the landscape (machine learning, computer vision, natural language processing, deep learning, neural networks, etc.), it’s not surprising that business leaders don’t have a clear understanding of how it will transform their businesses.

Also, the hype about AI is skewed. When new technologies hit the scene, evangelists and the media tend to focus on the opportunities and disregard the potential challenges. These skewed views fuel silver-bullet belief systems when silver bullets do not actually exist. It takes hands-on experience, including successes and failures, to truly understand the potential and limitations of a technology as applied to a specific business.

“It goes without saying that AI has the potential to completely transform business. I recently spoke on a panel about this topic at Fortune Global Forum in China, and everyone there, from prime ministers to chairmen of Fortune 500 firms, discussed the transformational potential of AI,” said Duffy. “There is a broad understanding that it is going to be transformational, but the challenge is that it requires work and investment to develop the strategy [and] vision to realize that potential.”

Many organizations are in the early stages of AI adoption, so they have not yet invested sufficient time and money in the process. In order to bridge this gap, leaders need to start gaining experience now, developing initial use cases or proofs of concept. Duffy recommends investing in a big-picture strategy, and developing a vision for how this could transform a firm or sector.

AI is more than a technology

AI is a piece of the digital transformation puzzle. As with all things related to digital transformation, technology is only part of the picture. The most effective strategies focus on business problem-solving.

“I believe companies will have the most impact with a business-first, value-led approach,” said Duffy. “The best way to approach AI is to focus on how to add value to a business beyond just cost efficiencies. Businesses must think now about AI from a strategic perspective and ask themselves how much more value they can deliver through more intelligent use of AI.”

Of course, there are some barriers to adoption that are technology-related. As is typical in the early adoption stages of a technology, the initial tools tend to be targeted at a narrow, technical audience that is capable of using them. However, as the technology matures, easier-to-use tools follow and abstract the some of the complexity. Usually those tools are aimed at “power users.”

Finally, becomes easy enough for the masses to take advantage of, such as analytics dashboards in the enterprise. Already, AI is built into and will be embedded in many kinds of devices and software, to the point where it is transparent to the user. For example, one does not have to be an AI expert to use Amazon Echo.

AI will create winners and losers in every industry,” said Duffy. “AI is here today and can provide significant value now. Can you really afford to be slower to adopt it than your competitors?”

Business leaders and organizations should get familiar with AI technology now, because it will make it easier to determine where AI can be used as an effective problem-solving solution, Duffy said.

Business leaders and technologists need to work together. EY does this internally to meld cultures and disrupt traditional ways of thinking.

Set reasonable ROI expectations

In the AI brief, Duffy and Mazzei say, “Many early projects will have low ROI and a limited impact.”  So, at what point, then, should businesses invest in AI?  On one hand, the early adopters gain insight and experience that those sitting on the sidelines miss. On the other hand, those who are later to the game have the luxury of using more mature toolsets and learning from others’ mistakes.

“Early adoption doesn’t necessarily have a low ROI. [To clarify,] the early adopters are often focused on the technology rather than the business problem – this can lead to low ROI,” Duffy said. “However, [early adoption] does lead to invaluable learning.”

Early technology-led projects may also have low ROI because they are (and should be) as much about learning as about value. Rather than limiting the scope to only technology-led projects; businesses should identify projects based on their business value and have them led by business stakeholders.

“Because of the transformational potential of AI, if you wait and your competitors don’t, you will be at a disadvantage. AI will differentiate between winners and losers, and the pace at which that is happening is only accelerating,” said Duffy. “Most people are early in their AI journey and the actual investment can be small relative to the potential. It’s a smart decision to make a relatively small investment to start.”

Overconfidence can be dangerous

The immense interest in AI is creating career opportunities and with it overstatements about qualifications. Duffy said it’s important to get the right talent.

“The Dunning–Kruger Effect is of significant concern in this space, that is, people can be unskilled and unaware of it,” said Duffy. “The field has grown so rapidly that there are many people who can solve technical problems, but they have a lack of deep experience.”

EY conducted a survey of 200 senior AI professionals, 56% of which said that a lack of talent is the greatest barrier to implementation within business operations. If companies don’t have competent AI professionals, they face three big risks that are easily preventable if business leaders think about them in advance and do something about them proactively.

The first is testing. It’s much easier to get AI testing wrong compared to other technologies. Getting it right requires a certain amount of sophistication, as there are many subtle statistical issues. According to Duffy, AI testing requires talent deep expertise, working with this type of technology.

[Maybe it’s time for your organization to make a real investment and commitment to AI. Read more here.]

The second challenge is that machine learning can amplify bias, which was another one of the key takeaways from the recent EY AI survey. Forty-one percent of the survey participants said they see the gender diversity of existing AI talent influencing machine biases. Researchers need to be especially mindful of bias, specifically, racial, gender or other cultural biases. To proactively avoid those types of bias, organizations will need to ensure that they’re hiring from a diverse talent pool when hiring AI talent.

Finally, AI tools are making automated decisions, quickly. A sophisticated monitoring system needs to be in place to ensure that anomalies are caught quickly, Duffy said.

How to ask the right questions

Business leaders who lack experience with AI may wonder how it’s possible to know whether they’re asking the right questions in the first place.

“Some of this is about building up experience over time, which reflects back to my point about how it’s better to be an early adopter. You start asking the questions, seeing the answers, and seeing the outcomes that lead to asking better questions,”‘ said Duffy. “By starting soon, leaders can get experience in determining how AI can have the most meaningful impact on their business.”

How To Increase Contributed Article Placement Success

abstract-1260505_640Pitching and placing contributed articles is a staple of a good PR program.  Editors are bombarded with ideas every day.  Some make the cut, some don’t.  Want to up your chances?  Think and write like a journalist.

I’ve received a few pitches lately that I found quite incredible.  I imagine the PR reps thought the pitches were logical – I certainly would have before I had journalism experience myself – but sitting in the chair of a journalist, I could see how faulty their strategy was.

The idea was, “How about if my client contributes an article that highlights the features of its product?”  My response was, “Sounds like a sponsored editorial product.”  Why?  Because the proposed content was essentially an ad.

My audiences – business executives and technologists – don’t have much of an appetite for blatantly self-promotional prose.  Moreover, promotional content does little to establish your client as “a thought leader.”  It positions them more like a salesperson.  The same can be said for other media including video and webinars.

The best contributed pieces really show off an expert’s chops.  That person knows more about leadership or emotional analytics or programming in Python than most of his or her peers do, and that person is willing to share their expertise, free of blatant product or service tie-backs.  A good pitch reflects that.

I bring this up because I hate to see people waste time and their clients’ budgets.  There are better and worse ways to do things.  It is entirely possible to advance your client’s agenda without attaching flashing lights to it.

Having said all of this, I see contributed articles published that are self-promotional, especially in the tech pubs have had their budgets slashed dramatically.  I don’t think those articles do the community or the contributor much good, so I’m sticking to my guns.

If you want to improve your chances of making it into the little pile, think and write like a journalist.



Deloitte: 5 Trends That Will Drive Machine Learning Adoption

Companies across industries are experimenting with and using machine learning, but the actual adoption rates are lower than it might be seem. According to a 2017 SAP Digital Transformation Study, fewer than 10% of 3,100 executives from small, medium and large companies said their organizations were investing in machine learning. That will change dramatically in the coming years, according to a new Deloitte report, because researchers and vendors are making progress in five key areas that may make machine learning more practical for businesses of all sizes.

1. Automating data science

There is a lot of debate about whether data scientists will or won’t be automated out of a job. It turns out that machines are far better at doing rote tasks faster and more reliably than humans, such as data wrangling.

“The automation of data science will likely be widely adopted and speak to this issue of the shortage of data scientists, so I think in the near term this could have a lot of impact,” said David Schatsky, managing director at Deloitte and one of the authors of Deloitte’s new report.

Industry analysts are bullish about the prospect of automating data science tasks, since data scientists can spend an inordinate amount of time collecting data and preparing it ready for analysis. For example, Gartner estimates that 40% of a data scientist’s job will be automated by 2020.

Data scientists aren’t so sure about that, and to be fair, few people, regardless of their position, have considered which parts of their job are ripe for automation.

2. Reducing the need for training data

Machine learning tends to require a lot of data. According to the Deloitte report, training a machine learning model might require millions of data elements. While machine learning requirements vary based on the use case, “acquiring and labeling data can be time-consuming and costly.”

One way to address that challenge is to use synthetic data. Using synthetic data, Deloitte was able to reduce the actual amount of data required for training by 80%. In other words, 20% of the data was actual data and the remaining 80% was synthetic data.

“How far we can go in reducing the need for training data has two kinds of question marks: How far can you reduce the need for training data and what characteristics of data are most likely minimized and which require massive datasets?” said Schatsky.

3. Accelerating training

Massive amounts of data and heavy computation can take considerable time. Chip manufacturers are addressing this issue with various types of chips, including GPUs and application-specific integrated circuits (ASICs). The end result is faster training of machine learning models.

“I have no doubt that with the new processor architectures, execution is going to get faster,” said Schatsky. “[The chips] are important and necessary, but not sufficient to drive significant adoption on their own.”

4. Explaining results

Many machine learning models spit out a result, but they don’t provide the reasoning behind the result. As Deloitte points out, business leaders often hesitate to place blind faith in a result that can’t be explained, and some regulations require an explanation.

In the future, we’ll likely see machine learning models that are more accurate and transparent, which should open the door for greater use in regulated industries.

[Deloitte also recently discussed 9 AI Benefits Enterprises Are Experiencing Today.]

“No one knows how far you can go yet in terms of making an arbitrary neural network-based model interpretable,” said Schatsky. “We could end up hitting some limits identifying a fairly narrow set of cases where you can turn a black box model into an open book for certain kinds of models and situations, but there will be other scenarios where they work well but you can’t use them in certain situations.”

5. Deploying locally

Right now, machine learning typically requires a lot of data and training can be time-consuming. All of that requires a lot of memory and a lot of processing power, more than mobile and smart sensors can handle, at least for now.

In its report, Deloitte points out there is research in this area too, some of which has reduced the size of models by an order of magnitude or more using compression.

The bottom line

Machine learning is having profound effects in different industries ranging from TV pilots to medical diagnoses. It seems somewhat magical and somewhat scary to the uninitiated, though the barriers to adoption are falling. As machine learning becomes more practical for mainstream use, more businesses will use it whether they realize it or not.

“[The five] things [we identified in the report] are converging to put machine learning on a path toward mainstream adoption,” said Schatsky.  “If companies have been sitting it out waiting for this to get easier and more relevant, they should sit up instead and start getting involved.”

What Data Analysts Want to See in 2018

The demand for data analysts is at an all-time high, but organizations don’t always get the value they expect, mainly because the organization, or parts of it, are getting in the way.

Being an analyst can be a frustrating job if your position isn’t getting what it needs in terms of data, tools and organizational support. Are you getting what you need? Here are some of the things your contemporaries are saying.

More Data

Despite the glut of data companies have, analysts don’t always get the data they need, often because the data owners are concerned about privacy, security, losing control of their data or some combination of those things.

“The problem of data ownership and data sharing is universal,” said Sam Ruchlewicz, director of Digital Strategy & Data Analytics at advertising, digital, PR and brand agency Warschawski. “For analytics professionals, these artificial barriers hinder the creation of comprehensive, whole-organization analyses that can provide real, tangible value and serve as a catalyst for the creation (and funding) of additional analytics programs.”

Jesse Tutt, program lead of the IT Analytics Center of Excellence at Alberta Health Services said getting access to the data he needs takes a lot of time because he has to work with the data repository owners to get their approval and then work with the technologists to get access to the systems. He also has to work with the vendors and the data repository subject matter experts.

“We’ve worked really hard getting access to the data sets, correlating the different datasets using correlation tables and cleaning up the data within the source systems,” he said. “If you ask a specific set or data repository what something is, it can tell you, but if you can snapshot it on a monthly basis you can see a trend. If you correlate that across other systems, you can find more value. In our case, the highest value is connecting the system and creating the capability in a data warehouse, reporting you can correlate across the systems.

Four years ago, people at Alberta Health Services wanted to see trend data instead of just snapshots, so one system was connected to another. Now, 60 connected data sources are connected with 60 more planned by the end of 2017. The company has a total of about 1,600 data sources, many of which will be connected in the next couple of years.

More Respect

The most effective data analytics align with business objectives, but what happens when your data analysts aren’t informed? Warschawski’s Ruchlewicz recently had dinner with the CEO of a large, international agency who spent millions of dollars on a marketing campaign that failed simply because the executive didn’t want to listen to “the analytics kids.” Never mind the fact that the analytics team had identified a major issue the target audience had with the client’s brand.

“[The CEO] dismissed them as analytics kids who didn’t know what they were talking about and proceeded to launch the campaign,” said Ruchlewicz. “Only later, after millions of dollars in spending (with no results to show for it), did the CEO allow them to make their case and implement their recommendations.”

Ultimately, their recommendations turned the campaign around. Ruchlewicz said.

“I wish this as a one-off story. It’s not. I wish this was confined to ‘old school’ companies. It’s not,” said Ruchlewicz. “Until analytics teams are given a seat at the table where decisions are made, analytics will continue to be undervalued and underappreciated across the entire organization.”

Analysts have to earn respect like anyone else, however. That requires communicating to business professionals in business terms.

“Executives and investors today are hyper-focused on the bottom line, and most that I’ve interacted with perceive analytics as a line item expenditure,” said Ruchlewicz. “[A]nalytics professionals need to take the first step toward resolution. There are several methods that allow the creation of a rigorous, defensible first approximation, which is sufficient to get the conversation started (and usually, some data shared).”

To help turn the tide, analytics practitioners are well-advised present information and construct business cases around their activities.

More Consistency

If everyone in the organization used the same terminology for everything, always had the right database fields accessible, and always entered data correctly and in the same manner, some enterprise data would be much cleaner than it is today. However, the problem doesn’t stop there

“If a person says, ‘I want an analytical tool,’ how do you group that and do trending on it when a person may call it one of the 100 different analytical tool names or they’ll say I need to do analysis on data? The words the submit are often different from what they actually want,” said Alberta Health Services’ Tutt

Tutt and his team are endeavoring to better understand what people are requesting in service desk tickets so the company can manage its software investments more effectively. Now that his team has access to the different systems, they know who’s using a product and when they used it. They’re looking at the problem from a Robotics Process Automation (RPA) perspective so software can be automatically removed if it hasn’t been used in a certain time period.

More Power to Affect Change

Industry analysts are pushing back on “data-driven” mantras because they think companies should be “insight-driven.” While they have a valid point, insights without action have little value.

For example, a large U.S. health provider has a massive analytics team that’s generating highly-actionable insights, but those insights are not being acted upon by the business. They can meet with a functional unit such as risk or compliance and show them insights. The operating unit will say, “That’s interesting,” but there’s no way to connect insights and action.

“The data teams are frustrated because they’re not getting the operational support they need,” said Adam Nathan, CEO and Founder of analytics strategy firm The Bartlett System. “The data teams don’t know how to drive that, except to get frustrated and quiet and get more value elsewhere. I think the tipping point will come when the company realizes it’s falling behind competitors. They’ll realize the company isn’t getting the value it could from analytics and that will put pressure on them to do something with those insights.”

How CIO/CFO Relationships Are Evolving

Digital transformation is driving huge organizational changes, not the least of which is the evolving relationships of CIOs and CFOs.

Traditionally, the two roles have been somewhat at odds because CIOs must continually invest in technologies and CFOs are ultimately responsible for financial performance. In today’s’ highly competitive business environment, CIOs and CFOs need to partner at a strategic level to drive growth and enable organizational agility.

From old school to new school

Data provider Dun & Bradstreet is going through a digital transformation that allows the 176-year-old company to behave and compete like a much younger entity. To get there, the CFO and former CIO (now Chief Content and Technology Officer) are working in partnership to set strategies and execute them.

“We come together quite a lot because what we’re trying to drive is more innovation at a faster clip in a more efficient way,” said Richard Veldran, CFO of Dun & Bradstreet. “It all comes down to data and technology which is at the core of many of the things we’re trying to get done here.”

As the sheer amount of data continues to grow exponentially, Dun & Bradstreet has more opportunities to drive growth by monetizing data. However, to do that, the CFO and CTO need to work as partners.

“So much of it now depends on the alignment of your technical capabilities and investments,” said Curtis Brown, chief content and technology officer at the firm. “Rich and I spend a lot more time talking about our strategy and our execution against that strategy. I would say that’s the single biggest change.”

The partnership allows Veldran and Brown to allocate resources more effectively and make joint decisions about where to invest and how to invest. They’re also working together in a lean agile fashion which enables them to accomplish more in less time while reducing the risk of big project failures.

Focused on high growth

Hitachi Vantara CIO Renee McKaskle and CFO Lori Varlas act as if they’re co-founders and, in a way, they are. Both women were hired into their respective positions about two years ago to spearhead digital transformation. Years before, Varlas and McKaskle had become acquainted while working at Peoplesoft.

“We’re two women in non-traditional women’s roles, so from the get-go, we bonded on the common vision of where we’re going to take this company and how our individual skills and experiences added to that story and towards that journey,” said Varlas. “I think the other thing that bonded us was time is not our friend, particularly in terms of technology, so we had quickly align on what the business strategy was and figure out how we leverage our own backgrounds and experiences to make that vision a reality.”

They both say it’s important to learn from each other, listen to each other and be aligned on the vision or outcome.

“As we work really closely with the business, things come up. Someone might approach Renee or [me] for different purposes, but it springs to mind, ‘Has Renee’s cybersecurity team looked at that?’ ” said Varlas. Or, “Does Lori know about that for investment purposes?” said McKaskle.” There’s a bit of a tag team going there because we both have a common understanding and purpose of how it fits together.”

Empathy is key

Cross-functional collaboration is necessary to drive effective digital transformation; however, everyone interviewed for this blog said empathy for the other person’s role is critical.

“I can sometimes be a propeller head, but to think more empathically and as a partnership toward the enablement and delivery of the operation of the company, that’s where folks sometimes get stuck,” said Dun & Bradstreet’s Brown. “CFOs do have to put pressure on delivering a certain set of results within a certain financial framework while [CIOs and] CTOs are trying to drive technical improvements that often require investment.”

As businesses undergo digital transformation, the CIO and CFO have to move quickly and in unison. The best results come when they’re aligned on the business outcomes they’ve trying to achieve. That alignment also helps CIOs and CFOs overcome some of the tensions that stem from traditionally separate roles.

Beware Analytics’ Mid-Life Crisis

Businesses are using analytics to stay competitive. One by one, departments are moving from static reports to modern analytics so they can fine-tune their operations. There’s no shortage of solutions designed for specific functions, such as marketing, sales, customer service and supply chain, most of which are available in SaaS form. So, when it’s possible just to pull out a credit card and get started with an application, why complicate things by involving IT?

Freedom from IT seems like a liberating concept until something goes wrong. When data isn’t available or the software doesn’t work as advertised, it becomes the IT department’s job to fix it.

“I used to call this the BI mid-life crisis. Usually about a year and a half or two years in, [departments] realize they can’t report accurately and then they need some help,” said Jen Underwood, founder of Impact Analytix, and a recognized analytics industry expert. “Now I’m seeing more IT involvement again.”

Organizations serious about competing on insights need to think holistically about how they’re approaching analytics and the role of IT. Disenfranchising IT from analytics may prove to be short-sighted. For example, a proof of concept may not scale well or the data required to answer a question might not be available.

Analytics’ long-term success depends on IT

IT was once the sole gatekeeper of technology, but as the pace of business has continued to accelerate, departments have become less tolerant of delays caused by IT. While it’s true no one understands departmental requirements better than the department itself, IT is better equipped to identify what could go wrong, technically speaking.

Even if a department owns and manages all of its data, at some point it will likely want to combine that data with other data, perhaps from a different group.

“We became accustomed to IT organizations managing the database architectures or the data stores and any of the enterprise wide user-facing applications,” said Steven Escaravage, vice president in Booz Allen Hamilton’sStrategic Innovation Group. “I think that’s changed over the last decade, where there’s been a greater focus on data governance, and so you also see IT organizations today managing the process and the systems used to govern data.”

Additionally, as more organizations start analyzing cross-functional data, it becomes apparent that the IT function is necessary.

“IT plays an important part in ensuring that these new and different kinds of data are in a platform or connected or integrated in a way that the business can use. That is the most important thing and something companies struggle with,” said Justin Honaman, a managing director in the Digital Technology Advisory at Accenture.

Where analytics talent resides varies greatly

There’s an ongoing debate about where analytics talent should reside in a business unit.  It’s common for departments to have their own business analysts, but data science teams, including data analysts, often reside in IT.

The argument in favor of a centralized analyst team is visibility across the organization, though domain-specific knowledge can be a problem. The argument in favor of decentralization is the reverse. Accenture’s Honoman said he’s seeing more adoption of the decentralized model in large companies.

Hybrid analytics teams, like hybrid IT, combines a center of excellence with dedicated departmental resources.

Hot analytics techs

Machine learning and AI are becoming popular features of analytics solutions. However, letting machine learning loose on dirty and biased data can lead to spurious results; the value of predictive and prescriptive analytics depends on their accuracy.

As machine learning-based applications become more in vogue, analytics success depends on “the quality of not just the data, but the metadata associated with it [that] we can use for tagging and annotation,” said Booz Allen Hamilton’s Escaravage “If IT is not handling all of that themselves, they’re insisting that groups have metadata management and data management capabilities.”

Meanwhile, the IoT is complicating IT ecosystems by adding more devices and edge analytics to the mix.  Edge analytics ensures that the enterprise can filter meaningful data out of the mind-boggling amount of data IoT devices can collect and generate.

In short, the analytical maturity of organizations can’t advance without IT’s involvement.

Just a Bit of Advice:  

Strategies for Successful Analytics

A few helpful hints as you move through your analytics journey.

If you’re just getting started on your data and analytics journey, think before you act.

Steven Escaravage of Booz Allen Hamilton noted, “I tell clients to take a step back before they invest millions of dollars.” Among other things, he said, make sure to have a good foundation around what questions you’re trying to solve today and the questions you perceive are coming down the path.

“Let’s put together a data wish list and compare it to where we’re at, because usually you’re going to have to make investments in generating data to answer questions effectively,” he added. All the other pieces about methods and techniques, tools and solutions follow these actions.

If you’re at the pilot stage, beware of scalability challenges.

“Very rarely for sophisticated analytic problems would I lean on a typical Python pilot deployment in production,” said Escaravage. “You’d typically move to something you knew could scale and wouldn’t become a bottleneck in the computational pipeline.”

If you’re in production, you may be analyzing all kinds of things, but are you measuring the effectiveness of your solutions, processes and outcomes? If not, you may not have the complete feedback loop you think you have.

« Older posts Newer posts »