Harvard Business Publishing

Tom, an excellent write-up on the value of BI. When I was in technology years ago, it was difficult to describe why investment in a data warehouse made good business sense, other than to describe that the information could provide insight into and help with decisions that either needed to be made, or to evaluate success of decisions that had already been made.

Now that I have moved from technology to nursing, and actually to the intersection of both, I believe that today's movement into electronic health records and the BI it has the potential to provide, could improve patient care and outcomes if used properly.

You have a good point here and hope to read more of your comments on the topic. In the meantime, and in healthcare, there is significant focus on measurable outcomes to improve patient care while at the same time reducing inefficiencies and costs.

Why not provide secure and confidential access to patient data and use it to evaluate potential health risks, adverse events, and use this "HI" (Healthcare Intelligence) to take steps to prevent them?

insightful post Tom. I think the BI industry is definitely at an inflection point. The days of providing the entire data and letting the user slice/dice data and then making the decision are over. BI needs to enable decision making more directly.

Excellent insight. I have spent too many frustrating hours trying to distill specific scenarios (use cases) for decision support dashboards. The problem seem to be that executives are not comfortable to commit to one or two at any given time in a fear of a political backlash. It is similar to everybody asking for transparency and accountability, but also are scared of being exposed at the same time. My personal opinion though is - bring it on! This is a lot less of a technology challenge, than an intellectual one.

I agree Gregory, human ego has always been in the way of peak performance, and in BI it is an epidemic.

Rather than seeing BI as a 'replacement' for human decision making it needs to be seen as a tool to ensure that all relevant information is made available to make better decisions, and to be able to project likely impacts from various options.

After working in CPM strategy for many years I find that the biggest barriers to performance are always individuals who fear that their expertise will be 'replaced'...and along with it their jobs. These individuals often harbor information behind barricades, creating silos of power that sap the true strength of an organization and prevent true collaboration.

Some mid-level managers are also afraid of exposing their weaknesses, preventing top performers in their teams from 'outshining' them.

BI is a great leveller. It is a capability that should be on every desktop - not just managers and execs. Empowering individuals through personal dashboards not only allows them to self manage, but to also see where their activities contribute to the strategic objectives of the organization. This is very satisfying and incredibly powerful.

Functional silos are broken down as each team can more readily identify how their actions [or inactions] impact the deliverables of another.

BI is no longer a nice to have - its an absolute essential tool to successful competition.

Just as in solving crime - they say follow the money.
In business - it's follow the logic.

Logic Evangelist

Very interesting article. The problem I have with the way the term 'Business Intelligence' is used in general, is that it is always limited to internal intelligence. Decision Making can better be based on what happens in the market (e.g. falling demand) than on the consequences of these market developments in your datawarehouse (e.g. falling sales).

In our company, we set up an automated (and award-winning) intelligence system that automatically scans developments in our (metals) market and sends it to us. We then use these developments for our own decision making (our Marketing Plan updates itself every day) and send it (tailormade) to 5000 customers that use it for their decision making. And better decisions of our customers are better for our business too.

Is there a market, in you opinion, for personal decision support software applications? Would people pay for tools which would introduce intelligence into their investment, spending or budgeting decisions?

Excellent insight on BI and decision making. Tons of data roam around, but are they useful? The BI industry has to reinvent the wheel and vent and customized the data needed by its clients.

I envy the economy of your expression, Tom. You capture a complex truth that has been unfolding for many years -- decades even -- a few paragraphs and a useful image.

Tom, Excellent insight, and I do agree a dramatic shift is needed as the value firms are receiving from current investments in BI is not, in most cases, paying off. Rallying around decisions to be made is a novel imputes to improving the focus of BI efforts and will work in many cases; however I'm not convinced that the decisions to be made are always evident. Such decisions may be hidden in the data and yet unknown to the firm. Using the Health Care example above it would seem obvious we know the decisions that need to be made, but given so little progress has been made at improving the productivity of processing patient information (and systematically analyzing it) I have to wonder if the decisions currently being made are right to start with. Perhaps the new goal business intelligence initiatives is to clarify what questions must to be asked and answered before the decisions to be made can be determined? Can the data help set the agenda?

Hi Tom,

Great post. As former head of marketing for Business Objects for nearly a decade one of my big obsessions was that BI vendors pitched "better decisions" but delivered "data."

In fact, in my new life on the unstructured side of the data divide, I think there's two things missing:

- Delivering what's needed to make decisions -- putting information, and the right information, in context

- Uniting both structured and unstructured information

For too long BI meant data which meant structured data. We need to leverage all of company's information for decision making and we need to support the decision making process: not just bury people in more reports and data.

Best,
Dave Kellogg

In my book on data architecture, one of the key lessons I hope to impart is to not put the cart before the horse. To this end, one chapter is entitled, “Data Warehousing – “We’re All Set! We Bought Informatica!”

Many organizations think that building a data warehouse is simply a matter of having a powerful ETL tool to bring all of the organization’s data into a central repository called the Data Warehouse. In reality, ETL is one of the last functions to take shape.

The very first order of business has to be to determine what the organization hopes to accomplish with the data warehouse. This, in turn, entails defining the business intelligence which will be the motivating force behind the DW initiative. And this requires

- Knowing your organization’s mission, vision, goals, and strategies
- Knowing your organization’s critical success factors
- Knowing your Basic Performance Indicators (BPIs), Key Results Indicators (KRIs), and Key Performance Indicators (KPIs)

Everything in the data warehouse, and everything about your organization's business intelligence builds on the foundation of the organization's mission, vision, and goals, and the critical success factors required to achieve them. The next foundational layer is the measurements used to determine how successfully the CSFs have been accounted for in the business model.

By determining the benefits you hope to obtain from your data warehouse, in terms of how it will support the critical success factors (and thus, the key performance indicators) before starting the design, you will stand a much better chance of producing something which can truly shape the business and promote inspired decision-making.

By the way, I have consulted at some companies where they had defined up to 200 KPIs and were trying to support all of them in their data warehouse initiative. This is a recipe for failure! You should have no more than 20 KPIs motivating the business intelligence coming from the data warehouse, and preferably 10-12. Don’t confuse basic performance indicators (BPI) with key performance indicators (KPI).

B I is fundamental for the development of third world economies, if those in charge of promoting such development could use B I to reach focussed decision that contribute to changes in the lives of people. we observe a tendency in those countries, in that there is Good prospect and policies in theory, but when it comes to the reality in life, it is the contrary that happens, it is sad to observe that the elite has fails the ordinary man in leading him to prosperity, a change is needed.

Regarding automated decisions we need to consider that BI systems are dealing with just the information that is "in the systems". Decisions take into account the information "in the systems" and supplement it with a lot of many other sources of information that are not in the systems.

I will be very suspicious of a decision that is based on just the information in the system. Untill we can address all the parameters and information sources for a decision we cannot make automated decisions.

Business Intelligence has to empower the people with the right information in the right format and in a timely manner. It is not to be thought of as some top level business KPIs - but people throughout the organisation utilizing the information in their activities. These business activities can be monitored and measured with the top level KPIs set by the BI systems.

Tom's article points in the right direction that BI needs to be aligned with the business goals at all levels.

Very interesting indeed. But I agree with Edwin Vlems that we need to expand our content horizon to make the best possible decisions.

Last year (2008) a new stream started in this context: Enterprise Information Management (EIM) which uses unstructured and structured content/data out of Enterprise Content Management Systems and CRM systems.

If we can complete this with external market information brought to us with some business mash ups tools like Kapow or Serena, we are going into the right direction in my opinion.

I do not think anyone in business would disagree with Tom's findings, with the exception of some BI software vendors who may have their own vested interests in having you believe otherwise.
I also do not see anything new in Tom's 10 principles of Business Intelligence.
The reality is that most businesses who own BI tools and software today use them mainly for glorified report development and do very little true analysis of data and information. Those that do slice-and-dice, drill down, mine and analyze will of course marry what they find with "structured human decisions".

I am disappointed with your posting. Several of the listed items are not ‘principles’ that would guide anyone to a better use of business intelligence; they are merely descriptions of your diagram. The diagram itself doesn’t really add any useful insight (or maybe I just don’t ‘get it’ from your description). Does it in any way help me to understand how business intelligence needs to change? I don’t think so.

The reporting and analysis function attached to business applications (by whatever marketing phrase has been applied over the years) has always been about supporting decision making, so I am not sure why you consider this to be ‘new’. Maybe it is simply a recognition that business intelligence systems and their operators have lost touch with why they exist. Time to push the big reset button and coin a new phrase (like ‘analytics’) to save us from our latest ‘legacy’ system.

Point 3 needs to be explained much better, because it could be interpreted as a case for restricting information flow – only provide decision makers with data you think they need to know. As a decision maker, I would rather make the ‘decision’ about what information is relevant and what is not. Information providers should not presume to try to make my life easier by telling me what is important (to them). Several comments allude to the presence of politics, but that is simply a reflection of the human element of decision making – numbers are not the only consideration when decisions affect people.

While I applaud your reminder that support of decision making is the end goal, and especially enjoyed seeing your criticism of ‘one version of the truth’ efforts, you are failing to recognize a key by-product of decision making: when decision are made, things frequently change. When the business changes, the old information is not always relevant, and new metrics may be required. Business intelligence efforts must continue to adapt with the business environment. I am surprised that your study did not reveal the need for more flexibility, faster response to business changes, or the ability to support new decisions every quarter. After all, if there is a need for something new, there must be something wrong with the old.

In response to Tom Davenport’s ‘The Next Big Thing’ and his 10 Principles of the ‘New BI’, I must apologize in advance, for saying some fairly harsh things, that will necessarily rub up against many ‘corporate sacred cows’ regarding BI. First, there appears nothing particularly ‘New’ about what Tom is saying here. I’ve been in the intelligence community for 30 years and in the corporate BI industry for 15 years, and yes I have seen it all before. Unfortunately I have seen it several times, and still nothing seems to have changed. It seems that each generation is destined to make the same mistakes, and have their starry eyed BI illusions dashed by reality of corporate practice.

My company has effectively given up on attempting to establish efficient BI in corporate environments. We tried to institute BI awareness by lecturing at Masters level at universities on MBA courses, and also by instituting BI establishment projects within the corporate sector. Both of these initiatives failed to ‘take’. The reason why was quite simple. There was a general lack of the necessary corporate CULTURE, that would support and nurture BI in the business environment. We now act solely as outsourced BI professionals to CEO who need a ‘drop in capability’ for a limited period. This is not ideal, as true BI needs continuity to maximize its power, but it does work and it produces quantifiable results.

Please pardon me if I react to the oh too predictable words and phrases that seem to follow BI about like a bad smell in the corporate sector. Words like:-
Data
Information
IT solutions
Datawarehouses
Marketing
IT systems
Dashboards
Data architecture
These trigger words are indicators that the wordsmiths have no understanding of what intelligence is about, let alone how to ‘get some in’! Mike Sullivan seems to have a fair handle on what he’s talking about, although I would take him to task when he used words like ‘providing decision makers with data’. WRONG WRONG WRONG!
..or decision makers demanding to know ALL when they are faced with making that decision. That Mike, smacks of a lack of trust in one’s BI professionals and seems to indicate that your ideal decision maker demands to micro-manage and indulge himself in business problem self analysis, rather than leaving it to his ‘A team’ BI staff. Overall I’d give you a strong B+ Mike and a possible A- to Paul Grill for his comment. Most of the rest of the comments are in the C- and F class, and sadly I have seen them all before. Some things seem to never change, and most seem to not be singing from the same song sheet.

Please indulge me gentle readers and allow me to make a few observations.

*BI is NOT about technology. It’s about what happens ‘between the ears’ of a skilled and trusted intelligence professional, working in lockstep with an enlightened CEO who trusts and is trusted. Truly effective and successful CEOs and BI professionals tend to come in matched sets! The BI professional CANNOT exist within the corporate power pyramid. He and his team need to operate outside it and they are answerable ONLY to the CEO.

*The basis of BI is the modified version of the hundred year old INTELLIGENCE CYCLE! You will not kind reader that the word cycle is used. That means that it an ongoing process! The cycle consists of Phases:-
1.The DIRECTIVE PHASE where the CEO and executive ask the specific ‘Intelligence Question’. (please ladies and gentlemen no corporate waffle here. This is where the CEO earns his stripes!)
2. The COLLECTION PHASE both internal, where much raw data, information, processed intelligence and corporate ‘wisdom’ resides and external input that needs to top up internal sources, with external data, and information, that are the grist for the ‘intelligence mill’, to be processed into relevant, usable and timely intelligence, supporting the CEO’s decision making demands, as set out in the Directive Phase.
3. ANALYSIS PHASE where raw data, information and even intelligence is collated, analyzed in context and processed into usable intelligence that will answer the CEO’s intelligence question.
4. DISSEMINATION PHASE where processed intelligence that will help the CEO to make decisions is presented in a form that is useful, relevant, in context and most importantly TIMELY.
5. REFLECTION PHASE, where the newly discovered intelligence is incorporated into Corporate WISDOM and from which flows ‘better questions’ and more focused intelligence tasking from the CEO and executive.

And so the CYCLE rotates once more, getting more focused and asking smarter questions. It really is as simple as that! Trouble is of course that we have to deal with ego, human personality failings, institutional politics, raw fear, game playing and a myriad of other ‘flies in the BI ointment’. Only a strong CEO can suppress these negative power that wreck the corporate intelligence cart.

I have a small problem with scenarios when it comes to BI. Sure there will be a few possible paths that seem to predict the future. Please remember kind reader that the future remains forever unformed. The best that one can do is establish the most probable courses and to stay attuned to developing reality and prune ‘the probability tree’ to suit, forever refining the probabilities of the future. It may be politic to mention at this time that the measure of a successful CEO is one who can make command decisions based in incomplete and occasionally contradictory intelligence. That is where the rubber meets the road in the CEO stakes!

Oh and by the bye. There is no such a thing as automated BI. It's a fiction. That reels of denial, abrogation of responsibility and pure laziness.

I could wax lyrical, but I trust that this snippet will give you a taste of where one should be looking when attempting to discuss BI.

Please no more discussions of technology, data and the like. It gets embarrassing after the first few times and indicates a total lack of understanding and leads back into the dead ends that BI ends up in about every decade!

Automated decisions do not mean better decisions. Once would program an algorithm to come up with a decision based on the input. This sounds more like automation than better decision making.

In my opinion the existing loosely coupled system is effective. The slice and dice model works if the business is patient and enthusiastic to discover. It is partly an issue with the business users not wanting to explore. Every end user would to like to look at data in a different manner. It is not practicable to build a huge report set to cater to everyone's needs. Rather providing broad subject areas and adhoc access to these subject ares might be a way forward.