Welcome to Chez Wherever, Featuring Our World-Class Data Science Chef!An analogy to relying on data scientists to do everythingPaul SimpsonBlockedUnblockFollowFollowingSep 16, 2017A data science team is the way to goFrom the plethora of news articles and blog posts about data analytics, more and more businesses are now keenly interested in having data scientists, and many now understand that they need a team of different data science types.
However, many organizations don’t realize the commitment they will need to make in order to get the most from their data analytics efforts.
Consider this analogy.
Imagine a company that has a typical cafeteria on premises, one that is rather “uninspiring” in terms of the food served.
However, the CEO of this company has recently become something of a bon vivant, a gourmet diner away from the office.
He now wants to see his company’s cafeteria use many of the same artful and amazing chef’s techniques to which he has been exposed at his favorite place, Chez Wherever (Why do I use italics? It simply makes the place appear fancier, but of course!), in the trendy part of downtown.
He talks up his idea to the board and they think it sounds pretty good.
The Chairman of the Board asks, “What do you need for that to happen?” But, the CEO doesn’t really know what all he would need.
He only knows how fine it looks, how exquisite the food tastes when it is brought to his table at the five star restaurant with the worldwide reputation.
But, there is one thing he does know, and by golly, that should be all that matters — that there is a Paris-trained gourmet chef at that restaurant.
And, so, the CEO offers a huge salary to hire him away from Chez Wherever.
And really, that’s all the CEO thinks he needs to do.
He doesn’t know that the kitchen at Chez Wherever is perfectly laid out and equipped for a chef to do his finest work (far better than the kitchen in his company cafeteria).
Nor is he aware that the chef was surrounded by a support team on each end, from the knowledgeable buyer of fresh produce (aka data engineers who can get the raw data from esoteric sources), and a kitchen staff specializing in tasks such as food preparation (ahem, data cleaning), a sous-chef (his understudy), and customer-facing serving staff members who excel at their jobs at Chez Wherever in the same way a good data analyst knows their way around Tableau dashboards and PowerPoint for sharing insights with the end users.
All were working in sync with the chef to ensure his creations are perfectly prepared and expertly served.
Sadly, the unfortunate chef has taken the new “company chef” job sight unseen, without thinking to inquire if the company is equipped with a support team and a kitchen resembling the kind to which he is accustomed.
You know where this is going.
He soon comes to regret leaving Chez Wherever.
The chef hasn’t got his effective tools, no support of the kind he once knew and is lacking the essential environment in which a true chef can do great work.
What This May Mean For Business LeadersI’m speaking here to business leaders who have decided it is time for their companies to jump into the maturity model for data-driven business decision-making.
They need to be educated on the infrastructure required to support effective data analytics, as well as the kinds of data team members who can support the work, on either end, of data scientists in the center of the data science workflow.
They need to know that it takes time, and stages of growth, and that they cannot simply hire a great data scientist or two away from Google or Microsoft and expect them to produce the same kind of analytics insights that will drive performance in ways the company has never dreamed of, and that will keep it competitive in the increasingly data-driven world.
The data scientist needs inputs — plenty of good quality data from an enterprise data management system (thanks to the advance work of data governance professionals and data stewards), and then handed off to them in an accessible format by experienced data engineers.
The data scientist ideally will then hand their product off to a UI/UX designer and software developers to productionalize the data science models in a clear, easy to use way that brings value to the business users.
And some data analysts to help explain the value to the business users.
…and What It May Mean for Data ScientistsThis also speaks a warning to present and future data scientists, who need to know that it is not practical to be a data science “team of one”, but rather that it is essential to have available and to rely upon the support players on both ends of their work.
For many of you, you have a challenge, should you decide to stay at your current companies that are striving to become data-driven.
The challenge is this — after you have gained data analytics education and experience, you have a need and a responsibility to inform and inspire the top managers of the value of data analytics and the need for the organization to rapidly mature in the areas of data management, data security, data ethics and data governance.
You can point them to articles such as this one.
This is essential if that organization has any hope of adequately leveraging your knowledge and skills as a data scientist.
An approach you can take is to educate and persuade the upper management in stages.
Stage 1 — Create Some EnthusiasmFirst, work up and present some infographics (or at least some dazzling data visualizations) that give a taste of what is possible, to try and get management excited about the idea of using descriptive, predictive and even prescriptive analytics to help the organization accomplish its goals.
This first stage is analogous to the introduction of personal computers into the workplace many years ago — their full potential was not understood by business people at first, and the data processing department hated the idea of a decentralized computer system, that they couldn’t control (a perceived threat to the dumb CRT terminals and mainframe programs).
No one knew (except the founders of Apple) just how these PC’s and Macs would transform many aspects of business, helping to automate dull tasks, streamline “word processing” for business correspondence, keep better records, introduce the power of spreadsheets and enable communication in new and effective ways, such as the ease of sending and receiving email, even between line-level workers and the heads of the company (which must have felt empowering), all of which we take for granted now, but which only became mainstream by the mid 1990's.
Likewise, when data science is introduced into the organization, it can be eye-opening for the management to learn how much more can be accomplished, how much more efficient and cost-saving the processes can become for identifying sales leads, high risk cases, fraudulent actors and supply chain management.
For example, they may find that a customer follow-up team now needs to work much less of a caseload, based on a new predictive model that shows them the smaller list of customers on which to focus their efforts first.
Remember that part about PC’s becoming mainstream?.It’s hard to believe the workplace was archaic and quaint only 20 to 25 years ago.
Likewise, data-driven organizations, I mean ALL organizations reaching the top couple of stages of the data analytics maturity model, will be the norm in less time than that.
I predict it will be in 10 years or less.
And then, people will look back to 2019 and say, “Did companies REALLY work in such an archaic and quaint way with their data back then?”Stage 2 — Explain and Persuade, Patiently, Again and AgainSecond, you will need to explain to the leaders why a more mature and integrated data infrastructure is needed, if they are to get the most out of their existing, and ever-growing data.
Beware, this will take time.
Trying to convince people that the old way needs to give way to the new way causes many to say, “no way”.
To make the best use of disparate data sources, there needs to be a common data area.
This may be a data warehouse (I can’t wait for them to go away, now that we have better options), a data lake or a data catalog — a well-networked collection of data sources, even unstructured ones, perhaps with APIs between them to translate.
Only in this way can an organization’s data be brought together into essentially one place to be used more fully by the analytics team’s tools and talents.
You must also persuade management to hire people who specialize in data management, people trained in data analytics and others to help get your work out to the users who need it.
Hopefully, this won’t all fall on you.
Executives are pretty sharp.
When they come across enough articles on this topic in Forbes and The Wall Street Journal, they will start investigating how their company can get in on the promising new technologies.
Go ahead, say to them the latest hyped-up buzz words as a mantra for them to repeat .
Is “Big Data” falling from its heights in the height curve yet?.Okay, replace it with “AI” or whatever current word or phrase is a close match to the systems your company needs to put in place.
Or, Perhaps It’s Time to Brush Up that ResuméAn alternative is to find a new job, one at a company that already fully or mostly understands what is needed for a healthy data science environment to grow and produce results.
You may need to find a company that is actively working to achieve a better infrastructure.
Now, back to the gourmet chef that our hypothetical CEO hired.
He has a choice to make.
He can either decide to continue at the company cafeteria in hopes of convincing the CEO that there will be no chef-quality delicacies possible until his working environment is changed to meet his needs, along with the right support staff, and the right food sources that are needed for his exquisite recipes.
Or, perhaps more logically, the chef can look around, and ask around, and eventually find a better-suited job at a place like Chez Somewhere Else.
After all, they already know what a chef does and are ready to welcome him into their culinary team.
Whether you are in the position of needing a data science chef or in the situation of becoming a data science chef, there are choices.
Here’s hoping for some well-prepared data science masterpieces in your near future!.