This article will also be published as an upcoming Analyst First’s white paper following feedback, authors are Eugene Dubossarsky and Jason Widjaja.
“Perhaps most importantly, analytical competitors will continue to find ways to outperform their competitors. They’ll get the best customers and charge them exactly the price that the customer is willing to pay for their product and service. They’ll have the most efficient and effective marketing campaigns and promotions. Their customer service will excel, and their customers will be loyal in return. Their supply chains will be ultra-efficient, and they’ll have neither excess inventory nor stock-outs. They’ll have the best people or the best players in the industry, and the employees will be evaluated and compensated based on their specific contributions. They’ll understand what nonfinancial processes and factors drive their financial performance, and they’ll be able to predict and diagnose problems before they become to problematic. They will make a lot of money, win a lot of games, or help solve the world’s most pressing problems. They will continue to lead us into the future.
– Competing on Analytics, Thomas H Davenport and Jeanne G. Harris
The notion of analytics as the basis of competition has been a point of discussion since the seminal HBR article ‘Competing on Analytics’ by Thomas H. Davenport in 2007. Its implicit premise is that that analytical competition is imminent – and that premise is sound based on the confluence of a surplus of data creation, increasing awareness of analytics as a strategy and scarcity in talent that makes effective analytics a worthwhile competitive edge. But both research and practice have been thin on the ground as to when this state of the world will reality.
This paper proposes that the answer to that question is two-fold:
- Firstly, that it will happen without much fanfare, because if something is truly delivering a competitive advantage – and analytics done well can certainly do that – it is in the organisation’s best interest not to let anyone know.It will be held under wraps as long as possible, with only its effects hitting the market. An ‘invisible arms race’ of sorts has already been underway for a few years, and what we read online is akin to the public reading scrubbed news of new offensives, techniques, and weapons in the field.
- And secondly, the short answer is ‘soon’.
Because winter is coming.
1 The Coming Winter Reckoning
“Oh, my sweet summer child,” Old Nan said quietly, “What do you know of fear?
Fear is for the winter, my little lord, when the snows fall a hundred feet
deep and the ice wind comes howling out of the north.
Winter encapsulates the state of business where stagnant and increasingly harsh operating conditions results in scarcity and drives cutthroat competition.
It is the state of the world where the conversation turns from ‘how much are we going to grow’ to ‘will we grow at all’. This effectively changes the conversation from the default posture of doing more of the same to a dire search for survival in the face of contraction.
When winter activates industry competition and reliable mechanisms of generating return start to fail, this is the time when doing you have always done is not enough. This is when decision making becomes crucial. And this is when analytics comes into its own.
Simply put, competing on analytics happens when growth and survival can only come at the expense of competition.
1.1 Analytics Comes Into Its Own In Winter
Cutthroat competition begets an emphasis on decision making, and decision making favours the agile. Taking a more structured look at organisational agility, this important capability can be viewed through the lens of OODA loop cycles of decision and response. The four dimensions of OODA are:
- Decide, and
Used in the context of competitive sense and response, the implication is that there are winners and losers in OODA. A holistic analytics capability and an analytics-friendly culture will bestow advantages across all dimensions of the OODA loop, and when organisations get caught in the OODA loop, the incentive to adopt analytics becomes readily apparent. As a side effect, winter also shines a light on analytics functions and punishes vanity and hollow analytics.
2 Winter Wartime: The Analytics Arms Race
‘You may not be interested in war, but war is interested in you.’ – Leon Trotsky
Winter brings the end of friendly competition. When the game changes from a positive sum game to a fight for a contracting pie, and your competitors are in the same boat, once ‘friendly’ competitors will turn increasingly hostile as they require a larger share of the market to maintain themselves.
What follows next is an arms race.
2.1 Once one industry player does analytics for real, everybody needs to do it for real
The nature of a well-functioning analytics capability is that it continually spots and exploits opportunity. Engaging internal decisions and external customers at both the most strategic and the most granular levels, a competitor’s only and best response is to answer in kind.
To an analytics observer, the signs and signals of arms race can be apparent. Teams scanning the horizon will see increasingly rapid changing circumstances. The unexpected happens more frequently as opportunities are systematically arbitraged by one or more players. Other signs may include:
- Predictive models failing increasingly short time frames.
- Strange pricing behaviour
- Different customer and product segments shrinking and growing
2.2 Analytical competition is limitless
Once analytical competition begins, it becomes a permanent feature of the landscape, because the scope of analytical competition is unbounded. Like competition among world class athletes, there is no ‘good enough’. What is good enough today may not be good enough tomorrow. This is the repeated game of competition.
This is also why the concept of ‘best practice’ is suspect – what applies to one context is often not transferrable, and to the extent that the details of a capability becomes publicly available as best practice (to be duly studied by the competition), it becomes questionable how much competitive advantage it actually bestows.
3 Outbreak is Imminent
Winter is bearing down on an increasing number of companies and industries. Is it also worth noting that analytics arms races have existed for years in our midst in industries where a predictive edge translates into direct and immediate value (such as hedge funds in the financial markets and the arrival of ‘moneyball’ in competitive sports), and the scythe of winter has rewarded and culled players in these industry. The same capabilities, pace and competitive dynamics that shape these industries spilling into new areas may herald the arrival of winter.
There are surfeit of trends that bear upon businesses more broadly and Australian business in particular. The following are selection we suspect are especially pertinent to ‘activating’ the coming era of analytic competition:
3.1 Businesses are in the throes of disruption
The threat of industry level disruption and new competitors providing better service at a lower cost and established industries coming under attack from new business models have driven major power shifts in Australia’s largest companies. Between 1994 and 2014, only one-third of the companies in the ASX100 are still on the index today. And of the ASX100 companies in 2014, one-third of them did not even exist in 1994. Through the increasing ubiquity of digitally enabled business models and start-ups, this trend is set to accelerate.
3.2 Disruptors compete and the basis of competition is shifting
In this transition from a world of IT scarcity to abundance, competitive advantage has little to do with unique access to technology, and everything to do with unique access to — and use of — information. When technology is near-ubiquitous, it’s the connection between people and information that drives business forward.
- HBR Jun 2013, How to Compete When IT Is Abundant, Aaron Levie
Our strategic and operating environments are awash with information technology and systems, and as they proliferate so too does the data that they produce. The consequence of the twin impacts of advanced technology deployments and an upsurge in data is that doing things right is getting easier, but doing the right things is getting harder.
Locally, other triggers may include macro-economic pressure from the end of the resource/mining boom and fallouts stemming from a precarious over-reliance on the Chinese economic engine.
All this ushers an era of competing for less.
4 The Good News: Preparing For Winter is Not Difficult
The advent of winter finds each organisation in various states of readiness, and its effects are already driving a wedge in relative organisational performance.
Regardless of where winter finds your business, all businesses can create an analytics function rapidly, cheaply and with a small footprint. This is borne out by data and practice – the arena is already seeing the rise of global analytics leaders, and a hallmark of their approach is their focus on iterative piloting of analytics efforts, while laggards adopt a wait-and-see approach or embark on large scale tool upgrades.
About Analyst First
Analyst First (A1) is a growing international organisation founded by practicing industry leaders with a central philosophy and a growing body of knowledge. A1 is about making analytics cheaper, more relevant and appropriate to organisations that can realise value from the application of analytics. It is also about presenting a radically different model of analytics to the one currently seen by most of the market.
 P19, Australia 2034: Luckier by design
Australia is now the most China-dependent economy in the world. Australian exports to China have grown from 8.5% of the total in 2003-04 to 32.5% in 2013-14 – and they continue to grow.
 AT Kearney, April 2014. Beyond Big: The Analytically Powered Organization
This article will also be published as an upcoming Analyst First’s white paper to be released shortly, authors are Eugene Dubossarsky and Jason Widjaja.
“Daunting as it may seem to rethink top-management roles and responsibilities, failing to do so, given the cross-cutting nature of many data-related opportunities, could well mean jeopardizing top- or bottom-line growth and opening the door to new competitors.”
Global spending on business analytics services is projected to rise from $51.6 billion in 2014 to $89.6 billion in 2018, according to a new forecast from International Data Corp. But how much of this rapidly growing, multi-billion dollar market is actually driving business value? And will the current executive leadership evolve in light of the rapid advancement towards a new business reality? This white paper explores both these questions.
1 Signs analytics is not driving strategic value
The gravitational pull of Big Data is now so strong that even people who haven’t a clue as to what it’s all about report that they’re running Big Data projects.
Gartner research shows that of the Top Big Data Challenges that organisations are facing, the largest by far is organisations are still struggling how to get value from big data.
And yet the prize is clear – a survey of executives at more than 400 companies around the world conducted by Bain & Company showed that the top 4% of companies who had the right people, tools, data and intentional focus significantly outperformed their peers. These companies are:
- Twice as likely to be in the top quartile of financial performance within their industries
- Three times more likely to execute decisions as intended
- Five times more likely to make decisions faster
The implication that analytics spending is preceding intention or understanding is the starting point of understanding why many analytics functions are operating at a sub-optimal level, and are a long way from realising the substantial value of possessing market leading analytics capabilities. The following are a number of signs that analytics is playing a peripheral role in many organisations.
1.1 Compliance vs. Real Analytics – Does Analytics Drive Competitive Decisions?
Analytics at its best creates value by providing decision support to advise decisions that have a material bearing on the company. These decisions can be both strategic – with analytics acting as the intelligence function of the business – or operational, with analytics adding value at key value-creating decision points within business processes.
However, considerable analytics work is primarily driven by regulatory compliance. This may be necessary, but it is ultimately a second order application of analytics – done not to edge out the competition or create strategic value, but because an organisation is obligated to.
1.2 Vanity vs. Real Analytics – Does Analytics Lead To Coherent Action?
The big data and analytics market will reach $125 billion worldwide in 2015, according to IDC.
While analytics budgets continue to expand at a brisk pace, many analytics functions are set up for failure in the absence of executive sponsorship, business adoption, links to organisational strategy and a clear mandate from senior management. Unfortunately, analytics capabilities are often started – even with significant budgets – without supporting structures or well defined business objectives.
This may be ‘vanity’ analytics at work – some executives may be driven more by the need to be seen to be taking action and keeping up with trends by having an analytics function, rather than actually leaning on analytics to drive a defined business outcome. A tell-tale sign that this ‘vanity’ analytics may be in play is the answer to the question “how often does the key sponsor or source of budget actually engage with the analytics?”
2 Four kinds of executives in the decision making and data landscape
The value of analytics can only be realised in context – it is dependent on the analytics function not existing in a vacuum, but being consumed in actual value-creating decisions. A discussion on analytics is incomplete without the related discussion about those who actually make decisions based on the outputs of analytics. Moving to a frame of thinking where analytics is embedded in day to day decisions both strategic and operational is not a trivial thing.
Not all executives are equal, and while it would be a rare senior manager to be unaware of the potential of analytics – anyone with an analytics budget would have been a vendor magnet for years by the time of writing – their actual comfort with data, willingness to engage with analytics and degree of ownership for de
cisions differ. The following diagram illustrates four broad categories of executives which outline the data and decision making landscape.
2.1 Type 1: Political Decision Evaders
The first category consists of executives who are primarily self-driven. While politics is an ever-present reality, Type 1 executives shy away from decision making before the fact in order to take credit for good decisions and distribute blame for bad ones, regardless of who actually made them. They encourage or abet a “stakeholder” model of decision making ownership where attribution of responsibility for decisions is unclear. At their core they lean on their savviness in organisational politics as opposed to fully owning and embracing their decision making authority.
A mark of such a decision evader may be they can be heard saying things like the following to avoid the responsibility of making decisions:
- “I’m just a facilitator”
- “It’s not my job to decide”
With respect to data, their interaction with data is largely cosmetic – they may invest in it for appearance sake, but not want to actually look at or engage with data.
As such, they will not add value in the more transparent and objective environment that is the hallmark of effective, appropriate and depoliticised analytics. The opposite may be true – the truth is inconvenient, so it does not always suit a political purpose. And effective, objective analytics may unceremoniously uncover their shenanigans, proving to be a threat to their continued existence.
Their talent at politics may allow them to survive, at least in the short term. However, their days are numbered.
2.2 Type 2: Earnest Skeptics
The bulk of ‘traditional’ decision makers, prevalent in large organisations and government are the group of managers who have the most to gain from harnessing the power of analytics. Doing the best they can using regular tools and techniques, when appropriately incentivised these Type 2 executives hold promise in the coming age of data.
They recognise that organisations need to use data, so they can and use data, and prefer to stay with what is comfortable, straightforward and widely used. Their typical arsenal includes an array of spreadsheets, operational reports and financials.
They recognise they cannot afford to not know which parts of the business make money and which do not, and engage with analytics accordingly. However, they may view “big data” as another of a string of fads and are rightfully sceptical of “big data” hype.
They are perhaps curious but unsure of the benefits of analytics, e.g.:
- What value does analytics deliver?
- How can they practice analytics appropriately to realise the value in their organisational context?
- How should their organisation instil that value in a cost effective way?
- How can they achieve, or start to achieve, that value with their current resources?
2.3 Type 3: Arms-Length Voyeurs
Type 3 executives are marked by their cursory engagement with data and their seemingly mild approach of being ‘willing to let data make decisions’. However, that facade belies possible organisational risks due to the mishandling of analytics. While slightly evolved in terms of their data and analytics maturity relative to Type 1 managers, they have yet to grasp that to leverage data effectively they must be prepared to change their own practices to fit its insights and discipline.
Unfortunately, their passive ownership may be synonymous with ineffective ownership. Investing in analytics is not a passive investment like a bank deposit, where an investor puts money away and expects it to give a return over time. It is rather an active investment more akin to a gym membership – its investors need to actively engage with what they get access to through the investment in order to get value out of it.
They want ‘actionable insights’ that tell them what to do. They think they are engaging with data and its insights, but they may not be due to two problems:
- By seeing analytics as a tool for actionable insight only, they are refusing to engage with analytics in a meaningful way. The limitations of analytics output need to be understood if they are to be used appropriately.
- By wanting the decisions they make to be automated, they are fundamentally devaluing the role that they do.
Type 3 executives may benefit from education in examining and synthesise complex information for decision making, or may ultimately opt to pare back their engagement with data.
2.4 Type 4: Engaged Pioneers
Type 4 executives represent the ideal sponsor for analytics and a model of analytics literate leadership who will propel their organisations forward through the intelligent application and embedding of analytics.
Possibly ‘analytically minded’ all along, even before the recent definition of analytics, they are executives whose essential role as true decision makers has not changed, but have been boosted through the additional, vital source of information that is data analytics.
In lifting these executives to fully realise the value of analytics, a significant amount of education and advice in the analytics field may still be required. However, it is even more important that they effectively organise, manage and cultivate their human and electronic analytics resources.
A Type 4 business leader makes intelligent demands of data analysts. Such a leader knows they need to be analytics literate. They also know they need to be educated about analytics in order to use it.
In embracing analytics, the work of a Type 4 executive is not necessarily made easier, but it does become more successful, and more fulfilling in terms of intellectual stimulation and organisational outcomes. By successfully orchestrating a well-functioning analytics capability to meet the ravenous demand for information from an analytically minded organisation, organisations will be able to:
- Make better decisions than their competitors
- Navigate existential risks
- Identify and exploit emerging opportunities
And it is these organisations that will be able to compete successfully and decisively break away away from their peers in the coming future.
Here is a recent presentation, a version of which went to Mark Burnard’s well-organised, catered and attended Big Data in Banking and Finance Meetup Group. Another version was presented to the Australian Superannuation Industry Trustee conference the following day. The Future of Analytics in Finance and Elsewhere, presented here in “director’s cut”/extended format contains much that didn’t make it into either presentation.
These are recent thoughts on what the future of data analytics looks like, what drives that future and how it is already here in many places, niches and organisations. Parts of this presentation will exapand into future blog posts in their own right. A video will also be forthcoming.
Speaking of videos: the long-overdue “Zen of Data Science” presentation, is up on the Presciient videos page (scroll down after “Zen of Predictive Modelling” to “Zen of Data Science” in 3 parts). These do deserve reposing on the A1 blog as its own article, to be done soon.
A good way to explain the most extreme yet pervasive issues in the “big data”/”analytics”/”data science” space is with parables. These transport those issues into other domains and thus hopefully help frame the problem.
The parable of “Big Letters” covers the issue of literacy, the willingness to “get on the bandwagon” without understanding or even being aware of some fundamental concepts, and the role that vendors can play in exploiting and maintaining this state of affairs. Here it is:
In 1439, Johannes Gutenberg first used the printing press in Europe, triggering the age of mass communication. An invention altered society. This was a time of revolution in the sciences, technology, economics and arts in Europe. A “Renaissance”.
The news reached a successful turnip seller in Europe. He saw enough conference presentations and heard from enough colleagues and others that printing was the information technology that was going to change business, and our farmer was determined not to be left behind. As it happens, the turnip seller, like so many at the time was in fact barely numerate and actually illiterate.
The turnip seller, consulted with printing press vendors who informed him of the different model printing press machines he could consider, and all the additional value he could how deliver. The vendors were not at all perturbed by the turnip seller’s innumeracy, and indeed did not even address it. After all, miracle though printing may be, it was still all about “technical”, “academic” things like reading and writing, and higher level things like “science” and “literature” in the same vein, hardly the domain of important business people like our turnip seller.
The vendors were very clear in their pitch: the turnip seller had to own a printing press, but in no way did they say anything about him having to do anything other than own it, follow prescribed “best practices” and, presumably, watch the money roll in. Learn to read ? Not mentioned, not relevant. Instead, the vendors presented some turnip-industry best-practice case studies, at which their presses excelled on a number of metrics. The case studies concerned key business applications, which all the other turnip sellers (competitors!) were getting in on. Could our seller afford to miss out?
Accordingly “Turnips – best practice” was printed, mass-produced in multiple languages. The seller watched his new printing team busily doing their jobs, and saw pamphlets, books and journals published. Naturally, he didn’t understand a word of it, but he knew that it wasn’t really his job to understand. What were all those words, letters, pictures for ? He didn’t need to worry. He was too important, “business focused”, “strategic” for that. Reading and writing was for the nerds that worked for him.
The press’ active life was short before it was long forgotten. It is unclear who was supposed to read it, or what value it generated. The seller found some value in showing his shiny technological marvel to friends, customers and colleagues, but never really understood what it was for. Soon, hard times befell the seller after a particularly harsh winter, and the printing press collected dust.
Sometimes, that winter, the seller did have time to wonder, what the heck do all those people the vendors cited as “best practitioners” do with their printing presses anyway.
This parable illustrates that spending money on “Big Letters”, or “Big Data” for that matter, before you understand what it can do, and its actual relevance to your business – is expensive, comical and a little insane. There is a minimum literacy requirement as well as a need for direct engagement with the technology and the insights it enables.
There is also a cautionary tale about “best practice”, and tool-centric, vendor driven procurement, with illiteracy on one end, and a sales model that isn’t exactly averse to that illiteracy on the other. Finally, consider under what circumstances the printing press vendor might have said “You know, maybe you don’t need a printing press, at least not yet. and I have no idea what you would do with it if you had one. In any case, you should learn to read first”.
Founder Stephen Samild presented some new ideas on Business Intelligence :
There may be a more detailed post on the subject by Stephen at a later stage.
Yours truly was interviewed by CeBIT. The topic was broadly : “What the heck is this Data Science Thing Anyway”.
“The Zen of Data Science” will be presented a second time, at the Singapore Analyst First group meeting today, headed by Brett Shadbolt who kindly continues to provide premises, refreshments, admin and many other kinds of support to A1, as well as his able leadership of the Singapore and Hong Kong chapters.
The presentation will happen about an hour after typing, so better hurry. Or wait until it presents in your town.
After an overlong absence from the blog, here is some material that should compensate at least in part.
These are three presentations that I have given over the last several months. First of all, a presentation to the Sydney Big Data Group all the way back in April.
The title was : “The Zen of Predictive Modelling”.
A Video of the talk
More recently, in June I presented to the Melbourne Users of R Network (MelbURN) on A1 related topics and the role of R in the A1 mix of ideas. Here is how it was billed on the MelbURN meetup page:
Looks like the title got your attention. This presentation is about R’s role, presence and profile in the world outside research and academia. Eugene will describe his experiences as a consultant, trainer, R User Group leader and general industry busybody in promoting the use of R and open source tools for analytics, in Australia and overseas.
He will discuss the political and social environment dominated by the brands of top vendors, and contrast this with a number of measures of R’s actual strong position, its prominence and indeed supremacy as most widely used analytics tool.
Eugene will also present a number of strategies for “selling” R, and pleasantly surprising clients with its power, graphical beauty and ease of implementation and use.
He will also discuss the role of Analyst First in this space.
Eugene may also discuss his new venture “ConnectR”, a community for crowdsourced, crowd funded and business-driven R development. Or he may go in any of a number of tangential directions. He usually does.
The talk pulled no punches, said some frank and less than kind things that needed saying, and promoted much debate during question time.
Finally, a presentation from a week ago, presented to the Sydney Data Miners meetup group, titled “The Zen of Data Science”
I met Eugene last night after connecting through the good people of Melb Uni Computer Engineering.
My career (I’m a 57 YO grey beard) included 10 years in Telstra where I was given a very broad brief to leverage scientific methods and analytical evidence in designing reforms of operations, and convince Execs to make changes. You’ll guess correctly I come from the operational strategy and change side of business. My disciplinary background includes a Ph.D in Management Accounting. I’m sure to stand in awe at the IT and mathematical competencies many of you will possess.
I’m looking forward to meeting many of you in your local chapter meetings. I’m in retirement number 2 – probably temporarily – and would love to use some of that time to learn from and, if I can, assist the good work of Analyst First and its individual members.
“Appropriate Empowerment” is the third and final element of the Holy Trinity, the three essential characteristics of sponsors of successful analytics practices, covered in the current series of posts. Appropriate Understanding and Appropriate Incentive were covered previously.
As before, this is an examination of the success mode and failure modes of the element in question. What does Appropriate Empowerment (or just “Empowerment” for short) look like when it succeeds, and what happens when it fails, or other elements fail to support it ? The success mode of Empowerment considers the situation where all elements of the Trinity are in place, but focuses on the role played by Empowerment.
The Failure Mode of Empowerment is the situation where the sponsor possesses Understanding and Incentive, but lacks Empowerment. We explore this situation, along with possible remedies, before concluding with the Isolation Mode, the situation where Empowerment is present, but alone, with neither Understanding nor Incentive in place beside it.
The success mode of Empowerment is simple, yet essential. Empowerment is the least visible element of the Trinity, more notable in its absence. Where the Sponsor sees the need for something to be done to the benefit of the business through analytics, and has the right Incentive to make it happen, then Appropriate Empowerment simply means : it happens. There is no one who can overrule, block, derail or otherwise unhelpfully modify any analytics initiative that has been put into motion.
Understanding ensures that the sponsor identifies the right analytics initiative for the greatest benefit to the business, and takes into account all that is required to enable it. Incentive ensures that the Sponsor actually wants this to happen. Empowerment then is simple : the Sponsor is in a position to launch the initiative, and ensure that it proceeds to the correct conclusion. He is able to support it with all the resources it requires, and protect it from unhelpful stakeholders. He is also in place to ensure that recipients of analytics recommendations act on them if the process requires them to do so. Tyrannical ? Perhaps. Far-fetched ? Certainly. But this is the ideal, however out of reach it may be for (current) real-world large organisations.
Empowerment is thus quite simple. It is the ability to make things happen.
It is also an absence of unhelpful constraints. A Sponsor with the Holy Trinity is sufficiently empowered not to worry about unreasonable or ill-defined expectations of value before the initiative or function is ready. Empowerment ensures that the function is not subject to IT-style management practices, deterministic waterfall and project management approaches. His analytics function is lean, agile and experimental : free to learn, fail repeatedly (for a time), as required to continually reach insights of massive value and exploit them.
The failure mode sets in when a sponsor has all the best intentions in terms of Incentive, and is well versed in Understanding what an analytics function can do, and what it requires to achieve it, alongside budget and a mandate to create the analytics function. Unfortunately, he may well lack the power to act as Understanding and Incentive may compel him to.
Any dilution of empowerment invites unreasonable expectations born of poorer Understanding and Incentive. A sponsor beholden to other managers, stakeholders etc is subject to constraints, expectations and pressures that may prevent an agile, experimental approach. The cargo cult of analytics, “Analytics in a Box” solutions promoted by some vendors stand in opposition to the agile approach, and enjoy attention and support from far too many senior executives. The resulting analytics cargo cult, subscribed to by much of senior management, expects great value from analytics, but does not know how to define this value, or even to measure it. This very lack of clarity may be what imposes inappropriate deterministic project management frameworks such as PRINCE2, and other inappropriate business analysis and management oversight by people who have no idea what they are managing or why. in such situations, project managers may grab the first objective metric, however irrelevant or minor and focus on it as a box ticking exercise. The analytics function is then little more than an IT production line, creating something of indeterminate value to satisfy a management fad. A sponsor beholden to such powers cannot be said to be sufficiently empowered. Worse yet, ignorant or indifferent management may relegate the sponsor under the auspices of IT. Needless to say, this is not an ideal outcome.
One large pathology crippling Empowerment is the modern corporate stakeholder model. A committee of stakeholders is not a Sponsor, especially when enough members of that committee have far from perfect Incentive or Understanding, and perhaps far too much Empowerment. A committee can be on the whole more stupid, poorly Incentivised and disempowered than any one member. A Sponsor beholden to such a Committee is hardly empowered, and the Committee as Sponsor is a far from ideal scenario. The fact that this situation is reality in so many large organisations does not diminish the fact that it is utterly pathological.
In the ideal situation the Sponsor is beholden to no one with excessive power who is inadequate in the other two key characteristics. The ideal Sponsor is therefore the CEO, and better yet a manager / owner. Again, this is perhaps unrealistic, but still needs to be identified as the ideal, and any deviation from it analysed in terms of potential failure of Empowerment. It is also the reason that the most innovative, valuable and agile analytics exist in tech startups and not large “Enterprises” (in quotes because they are usually the very opposite of that word)
Not all pathologies of Empowerment concern levels of power above the Sponsor. Other pathologies of Empowerment are lateral. The most immediate lateral power issue is one with IT : too many IT functions find it their job to block analytics access to tools, especially open source tools that are otherwise readily available, free and powerful. They may prevent access to adequate, and otherwise cheap and readily available hardware and useful online services such as cloud computing. They are also known for starving the analytics function of data. Too many analytics functions are in a situation where the main expenditure of effort is building business cases for data, tools and hardware. A sponsor who knows this to be the case but cannot fix it is clearly not sufficiently empowered.
Lateral Empowerment is also an issue with “trigger pullers”, people whose job it is to act on the recommendations of operational analytics. The most striking case of this is a pathology i have seen in a multitude of organisations making use of predictive operational risk analytics. Predictive models provide lists of targets (eg revenue leakage, non-compliance, suspicious behaviour, fraud risk indicators etc). In all cases a human being is provided a list of targets generated by the predictive model. Ideally, this human being proceeds to manually investigate the targeted cases. Unfortunately, in most situations, these individuals do not understand or trust these predictive models. In my experience, many such individuals cannot conceive the very idea of the inference of a model from data. It would appear that there are whole cultures of people who cannot imagine such a thing as statistical induction. They naturally voice their displeasure and challenge, stall and undermine the process. Much of a Sponsors job seems to be the thankless, draining and often never ending task of “winning them over”. A sufficiently empowered Sponsor would, however, be in a different situation. When asked why these people should trust these models he would be able to answer “because if you do not, I will fire you and perhaps hire someone who does what they are told. Or replace you with a smart pattern matching algorithm”. Again, this is perhaps not realistic, and perhaps suggesting something that certain Public Sector Unions would consider on par with a crime against humanity – asking that people do their jobs. The whole issue of uncooperative “trigger pullers” was only raised to make a point about Appropriate Empowerment: if a Sponsor is not able to ensure that human components of an operational analytics value chain do cooperate and act as a part of the analytics value chain, there is a failure of Empowerment. Perhaps effective analytics sponsorship, as defined in this series is impossible in most organisations where employee non-compliance and stakeholding is a given.
A lack of Empowerment is however, far from the end of the world, and the relatively dystopian situation described above matches many existing analytics functions, particularly in government and quasi-government organisations. They still manage to survive, and add some value, although arguably but a fraction of what would be possible if only sponsors were more Empowered. These organisations have in fact found themselves innovating in a number of fronts, dealing with insufficient Empowerment, and in some cases developing methods of generating more of it.
One key solution to the problem of insufficient Empowerment is Separation from IT. As far as possible, as quickly as possible, it is important to establish a “sandpit environment”, separate from the main IT network, where new hardware may be added, and software loaded outside of IT governance. This is essential if appropriate computational power and open source tools are to be leveraged quickly and effectively.
Another part of the solution, and one that is even more fundamental is Stealth Mode. It is imperative that a new analytics function has the ability to learn, experiment, and fail in its early stages. Expensive budget items such as vendor tools create massive, thought ill-defined expectations. Expectation management is yet another reason to avoid expensive vendor software early in the creation of an analytics function.
Ideally, the function has a small crew of capable, flexible people, a small budget and access to data and open source tools. Also, the function has a main focus that is a well-defined, business as usual task such as reporting. Actual analytics can be done on the side, as a side project, and not announced until it yields results. These results can then be presented as wins to formalise and Empower the nascent analytics function. There may then be sufficent leverage to acquire more staff, create a sandpit environment and acquire data reliably.
As discussed previously, the most important element of the Trinity is Incentive. With Incentive alone, the Sponsor knows that their first task is to increase their Understanding. Some of this is reading/study, some of this is consultation with experts, and much of this is experience which can be obtained in stealth mode. Empowerment is important, but as we can see it comes third in importance.
Indeed, most capable analytics professionals find themselves working for under-empower sponsors. This is not ideal, but not a career-ending situation. Indeed, the struggle for further Empowerment of the Sponsor is the defacto KPI of most analytics functions, and many professionals find it as exhilarating as they may find it frustrating.
It remains to discuss the “Isolation mode” of Empowerment. What happens when the Sponsor has all the power, but no Understanding, and,lacks the right Incentive? Here ignorance conspires with either a lack of real enthusiasm for Analytics, or an entirely different agenda, and gives them a hefty cheque book. So, what can happen ? A storm of Cargo Cults, management fads and buzzwords. “Analytics”, having something to do with “data” and software must clearly be some kind of IT, best managed and bought by the CIO and best explained by people who sell software. And that’s how the wrong kinds of Vendors happen. Long sales lunches. Exciting pre-sales presentations. Use of the words “Enterprise”, “Innovation” and “Insight” by people who don’t have anything to do with any of them. “Case studies” of previous such exchanges in other high profile corporations, presented as success. People who may not really care what they are selling, sell to people who don’t really understanding or care what they are buying. Consultants, the “best practice”, “brand recognition” kind jump in. More money gets spent. Everybody involved wins, except the (theoretical and distant) shareholders, citizens and other ultimate beneficiaries of the business. Almost always, none of the parties is an actual owner of the business in question. Most owners are far more sensible than that.
So what happens after that? Software get installed. Systems get integrated. People get hired, maybe, as an afterthought to mind the (far more important) Machines. These people are likely software developers, data base managers and project managers. Maybe even a token statistician. Gannt charts get ticked. Bonuses get paid (at least on the vendor side). Conferences benefit from new “Best practice” case studies. The Vendor-Consulting complex marches on in all its dinosauric grandeur.
So Incentive and Understanding matter, and Empowerment on its own is not a great idea, however common this situation may be.
About usAnalyst First is a new approach to analytics, where tools take a far less important place than the people who perform, manage, request and envision analytics, while analytics is seen as a non-repetitive, exploratory and creative process where the outcome is not known at the start, and only a fraction of efforts are expected to result in success. This is in contrast with a common perception of analytics as IT and process.
Tags in a CloudAIPIO analyst first Analyst First Chapters analytics analytics is not IT arms race environments big data business analytics business intelligence cargo cults collective forecasting commodity and open source tools complexity data decision automation decision support educated buyer EMC-greenplum forecasting HBR holy trinity human infrastructure incentives intelligence model of analytics investing in data lean startup literacy management culture MBAnalytics operational analytics organisational-political considerations Philip Russom Philip Tetlock prediction markets presales R Robin Hanson Strategic Analytics tacit data TDWI Tom Davenport uncertainty uneducated buyer vendors why analyst first