Guidance and results for the data-rich, yet information-poor.

Layered Pyramid Three LevelsHow long before we actually see some results here?

It’s a question the TMA team gets during a great many Q&A sessions. Everyone wants a time frame–some idea of the amount of time it takes to start getting actionable intelligence out of any given model.

But much depends upon the level of modeling that you are trying to engage in. There are three levels that you can choose from.

The first is the single model level. At this level, you’ll be building a single model for a single purpose. When starting from scratch, you can generally achieve this level within six weeks or so. However, in truth the single model level is not likely to produce much ROI. That’s not to say it’s useless–this sort of “level 1 model” makes an excellent pilot project, which can be used to demonstrate the benefits of predictive modeling to an otherwise recalcitrant leadership team.

The next is to build a repeatable process at the departmental level. This would be a model that’s run and updated on a regular basis. For example, your marketing department might regularly test the response to different advertisements in different zip codes to find out where it’s best to spend money and effort. They might re-run the model every six months to account for changes in external circumstances and demographics, and to adjust as the model prompts them to adjust. You can build a model like this in about six weeks, and you’ll get some solid use out of it.

The third and final level takes the most time–usually about six months. This is the “practice” level. That is, you’re going to create an analytics practice, a full department that serves the entire enterprise, a department capable of tackling a variety of business problems. As you can imagine, this is the analytics level that produces the greatest amount of ROI over the long run, even if you have to wait longer to see results.

One of the ways that TMA assists clients is by helping them build a true modeling practice, mentoring them so that the practice will ultimately be successful in the long term. It’s a great way to ensure that you maximize the actionable, profitable insights locked up inside of your data.

Remember, any grad student can use modern software to build a model in two weeks. Unfortunately, two weeks isn’t enough time to lay groundwork, provide context, or derive value.

It’s really only enough time to build a really fast rocket with no mission plan.

EverybodyDoingItAn inflammatory headline? Maybe, maybe not.

Think about it.

Everybody talks about big data.

Nobody really knows how to do it.

Everybody thinks everyone else is doing it.

So, everybody claims they are doing it!

In reality, less than 20% of organizations have begun to develop data projects. Those who have completed these projects successfully are in the minority. Even then, many of these organizations are unsure whether to treat that successful project as a spot effort, or whether to make it an enterprisewide concern.

That’s a shame, because data projects do offer significant value. For example, a recent study called How Does Data-Driven Decisionmaking Affect Firm Performance? indicates that, all other factors being equal, organizations who adopt data-driven decision making see 5-6% more output and productivity than organizations who do not adopt data driven decision making.

The disconnect occurs precisely because companies get into a “big data mindset” instead of adopting a mindset which focuses on deriving actionable insight about pressing business problems through predictive analytics.

In truth, “big data” is just an amorphous term. One TMA trainer defines it as, “one byte more than I can efficiently manage in the time allotted.” Thus, what looks like big data today is, in fact, going to look like small data tomorrow. Big data isn’t revolutionary–it’s evolutionary. The business world has the opportunity to manage data more successfully by choosing to realize that.

Keep an eye out for TMA’s newest seminar, “Big Data, Rhetoric and Reality,” if you are interested in exploring these issues a bit further. It’s coming soon, and it will help you navigate your way through a mess of hype so that you can truly extract real value from your organization’s data.

Composition with hardcover books in the library“Can you recommend some good books?” It’s a question that comes up quite often in TMA webinars.

Finding good books can be pretty difficult. That is because most of the books that are out there right now will actually lead you in the wrong direction.

This is because many books are still focused where many technicians are focused: on building better algorithms. They completely ignore the strategic approach that TMA teaches.

There is one good book. Competing on Analytics: The New Science of Winning” by Tom Davenport is a book which really focuses on getting the business value out of the analytics.

Of course, after you’re done reading Tom’s excellent work you might well want to dive in even deeper. At that point, it’s difficult to steer you in the direction of any good reading material.

That doesn’t mean you have to end your explorations, however. You can, for example, choose to download some open source software packages. Then, take a look at the documentation. Look at the example data sets.

This will give you a feel for how this type of analytics works within the context of setting clear, concise business objectives.

Of course, you can also choose to expand your education by signing up for TMA’s training courses as well. You’ll receive hours of materials which will help you dive into analytics in a way that will continue to produce results.

This is where you want to focus your energy anyway. You can build the most beautiful model in the world, but it’s not going to help you if it measures all of the wrong things. And reading all of the books in the world won’t help you until you wrap your head around a process which will help you derive actionable intelligence from the mountains of data that your organization has collected.

PresentationThis is a question that TWA hears a lot in the Q&A sessions held after each session of the free webinar. Usually, the person who is asking the question is already quite on board with the idea of using the insights that data mining can deliver, but his supervisors aren’t so sure that they want to invest the resources into any predictive analytics projects.

If you’re in this position, take heart. At the very least, it’s a sure bet that data mining and predictive analytics are at least on the radar of corporate leadership. Everyone’s hearing about “big data” and how data mining is being used successfully to create real business results. In fact, you can even see their caution as a positive. Too many management teams are all too “on board” too quickly, and so dive in with both feet without any plan or sense of what they are doing. The result is a lot of wasted time and money that does not drive actionable intelligence.

Your management team’s caution gives you the ability not only to sell them on the benefits of data mining and predictive analytics but to sell them on the benefits of launching these projects in the right way.

The Best Way

The absolute best way to get management on-board would be to contact TMA to schedule a day-long visit with seasoned experts. The first half of the day would be spent on an in-depth presentation while the second half would be spent on a round table where stakeholders would enjoy an open forum where they might air their questions and concerns.

Pilot Projects

If you don’t have the budget or the power to schedule a presentation then a pilot project becomes the next logical step. You’ll need an experienced model biulder to make the project as successful as possible.

Remember that you’re generally not going to get the chance to perform a full, comprehensive assessment of organizational and departmental goals as you launch your pilot. Nobody’s sold on using predictive analytics yet, and you will likely have other needs, goals, and projects tugging at your attention.

That means your results won’t be the best possible results. This is okay. You really only need results that are impressive enough to demonstrate the strength of the effort.

In order to avoid wasting too much time and/or money you should choose a project that looks likely to provide the most benefit with the least amount of risk.

What to Avoid

Avoid vendors.

Vendors may well sell your leadership on the benefits of data mining and predictive analytics. Vendors, unfortunately, are also there to sell software.

This is not to say that software is never going to be necessary. However, purchasing software at this stage is like purchasing a luxury sports car before learning how to drive. Software is not a magic bullet, and organizations who invest in software first tend to get caught up in hype. This means it becomes difficult to extract the value from the data, which puts companies right back into the position of being information rich but knowledge poor.

TMA is vendor-neutral for this reason. Our goal is to get you trained without pushing any particular agenda.

Most pilot projects can be completed with Excel. You don’t require better software so much as you require the ability to think critically about the problem and the organization’s goals. Show results, get the buy-in. You can always get the budget for a fantastic and needful software package later.

Looking for a data scientist is a little like going on a unicorn hunt.

Looking for a data scientist is a little like going on a unicorn hunt.

Are you trying to hire a “data scientist” for your organization? You might want to think twice before you decide to place that job ad.

“Data scientist” is either a meaningless designation or a descriptor for a person who will prevent your organization from realizing the full value of the data that it currently owns. Here’s why.

Data scientists typically approach the problem from the wrong direction.

Most “data scientists” typically focus on building technically superior models. There’s nothing wrong with building a better rocket ship, but first you’d best make sure that the rocket is actually pointed in the right direction.

No “optimized model” has ever aligned with business objectives. No business has ever generated a significant benefit from merely building a better algorithm.

In fact, many so-called “data scientists” pooh-pooh strategic assessment and project planning as “fluff” that distracts them from the “real work” of writing ever-more complicated code.

Unfortunately, strategic assessment and project planning happen to be vital if you’re ever going to extract any value from your data.

The term (as most organizations use it) describes a unicorn.

It is impossible–or at least, exceedingly rare–to find all of the skill-sets of a so called “data scientist” as most companies envision the position within a single human being. When organizations talk about “data scientists” they typically mean someone who:

  • Has a collection of advanced analytical skills
  • Has vast IT experience
  • Has and effectively uses a broad range of managerial soft skills
  • Can oversee analytic processes at the project level.

This mythical human somehow has managerial acumen and technical skill all rolled up into one brilliant, convenient package. Someone like this might exist in the sense that anything is possible…kind of like the way unicorns might exist in a universe where anything is possible. Since most organizations don’t have the time or money to embark on a unicorn hunt it’s smarter to take a step back and to think about who or what can actually achieve what the organization hopes to achieve by hiring a “data scientist” in the first place.

Anyone can call themselves a data scientist.

Granted, if you are dead set on acquiring a certain skill set then it’s awfully hard to fake having the technical skills. However, there is simply no formal definition for the term, which means no certifications, no degrees, and no quality controls. An unemployed MBA can legally hang out his “data scientist” shingle tomorrow. Often, amateurs do just that, to the detriment of the organizations they attempt to help.

Your existing employees can probably give you what you need.

Believe it or not, your existing employees probably have what it takes to help you derive outstanding value from the data you possess. In fact, training strategic thinkers who are close to the problems your organization is facing is often the first step. Sending key employees to a vendor-neutral training regimen which takes just a few short weeks can help you  begin transforming your data into actionable intelligence that offers solid benefits to your business. Doesn’t that sound far better than hiring an overpriced theoretical analytic specialist who is largely incapable of taking your organization where it needs to go?