HFTP News
September 28, 20206 min read

Citizen Data Scientists: In Search of Return on Intelligence

Article

By Sanjay Nadkarni, Ph.D.

The Emirates Academy of Hospitality Management

Should a café hire a barista or the engineer who designs the coffee machines? To brew a great cup of coffee, you need to know your beans and blends, the water type, how to operate the coffee maker and little else more. So, it's a no brainer who needs to get hired. While this is not to say that the engineer cannot be a great barista, the point is that when it comes to frontline delivery, domain expertise trumps the technical nuts and bolts knowhow. And yet, even as industries, including hospitality, increasingly flirt with adopting data science and all that is associated with it such as AI, machine learning, deep learning and the like (for brevity, let's just call it AI/ML), the dominance of data scientists, techies et al from build to run is pretty much evident. For the build part, yes, just like the engineer who designs the coffee machine, the data scientist has and needs to play the part. But when it comes to operations, the business side should be calling the shots. Yet it is usually the geeks who overwhelmingly run the show, ostensibly because the platforms/systems (designed by guess who?) are too complex for ordinary analytically challenged folk to use.

Who Makes the Decisions — Engineer or Frontline Operators?

In a "data rich/analysis poor" vertical like hospitality, where the primary domain expertise emphasis is on delivering guest experience, the attempts at adopting seemingly fancy tools like AI/ML get that much more awkward. Legacy IT is another story, which we will leave for some other time. Maybe it is an exaggeration, but one could uncork the bubbly if most hospitality decision-makers were able to differentiate a web service or solution from an algorithm-driven AI/ML approach. And I do hope this statement is a false positive.

The point is that data scientists who are experts in the AI/ML gobbledygook stuff usually tend to have little or no domain knowledge in hospitality and allied verticals. And if and when these types get hired to solve hospitality's pressing problems simply because hoteliers are too busy or scared to do anything meaningful with their data assets, this can be a recipe for disaster. The overall data science adoption failure rates over the past three years are summarized in Exhibit 1 below and the situation in hospitality is likely to be much worse. Going back to our café story, it is like the barista finding the coffee machine too complex to operate, compelling the cafe owner to hire the engineer who designed it to brew and serve the coffee. This is a pay more, get less scenario, since it is reasonable to assume coffee machine design engineers as a species are more rare than baristas, and yet may not quite be able to differentiate their Arabica from Robusta. So, are data scientists useless? Absolutely not, as without them, there is no AI/ML to begin with. But when it comes to data-driven decision making, the business side specialists need to be in the driver's seat.

60% of Data Science projects fail - Gartner (2016)
Fail rate on Data Science projects is closer to 85% - Gartner (2017)
87% of Data Science & Machine Learning efforts fail and never impact business - VentureBeat (2019)

Exhibit 1: Data Science Fail Rate Trends

Introducing the Citizen Data Scientist

Data are the oxygen for AI/ML and in this era of big data (volume, velocity, variety and the lot), there is no dearth of this oxygen. The data landscape in hospitality continues to get even more exciting with the overlay of IoT/sensor generated data on traditional data assets. Forecasts based on stale information and assumptions are passé; predictive, prescriptive, real-time analytics (PPRTA) for data-driven decision-making is where the action is. Dynamic pricing of which OTAs are the masters is only a manifestation of PPRTA and hotels need to up the ante by getting into this game, on and off premise. Well, easier said than done, right? How would hotels be able to afford those P.hD.s in data science, particularly with all this post Covid19 distress? And even if they could afford to hire such an expert, which data scientist is worth his or her salt or coffee beans aspire to a career track in the perceived backwaters of hospitality?

And this is where the citizen data scientist armed with autoML tools can come to the rescue. And who is a citizen data scientist? Someone whose primary job role is anything but data science, yet is able to deploy predictive and prescriptive analytics to inform his/her primary job role. And without knowing the math or having to write a single line of code, how can this be possible? AutoML (short for automated machine learning) wherein the end-to-end application of algorithms to real world problems right from data cleaning and preparation to model deployment is automated and doable in an intuitive non-code visual interface — this is what the citizen data scientist can use for data driven decision making. In other words, you, as the domain specialist in hospitality management and operations equipped with the right autoML tools, will be able to leverage the hotel's data assets way more productively than ever before. Not exactly the holy grail, but a pot of aromatic brew, nonetheless, to optimize return on Intelligence — the new ROI. The barista finally has a coffee machine simple enough to operate.

Sanjay Nadkarni, Ph.D. is director of innovation and research at the The Emirates Academy of Hospitality Management located in Dubai, UAE. His domain interests are in the convergence space of analytics, digital and sustainability in the services sector. Nadkarni oversees the HFTP Research Center - Middle East. 

HFTP

7301 Ranch Rd 620 N, Ste 155-193
Austin, TX 78726-4537
United States

Phone: 1 (512) 249-5333
www.hftp.org