BI Travel projects need THIS to be profitable

Large corporations discovered many years ago that outsourcing is a winning bet. A proven fact, magnified when it comes to software development. The IT universe is so vast, that it makes no economical sense to try and develop in-house all the possible solutions needed. Besides, “intrapreneurship” and “innovation” are top buzzwords these days, but the real thing is out of reach for slow-moving, Jurassic behemoths. As a consequence, industrial giants compensate by buying specialized, fast and lean startups (Hello, Amadeus!). Others try to modernize their aura by absorbing startups via acceleration programs, incubators, etc. But what about intermediate and smaller players? Is all the above valid about business intelligence solutions?
It’s no secret that, in the travel arena, technology is either invented or assimilated after years of delay, compared to other industries. BI is a term that has been around for decades in several environments, although popping out recently among hoteliers, distributors and tour operators. No surprise there: business was good for so long, it was unnecessary to gather intelligence. Only when balance sheets are grim, actions will be taken… although technology improvements would be the last ones considered. A case in point is BI: beloved but underutilized in large (and lucrative) corporations, a difficult to operate lifesaver for small organizations.
Regardless of internal policies or rooted beliefs, it is undisputed that BI can actually save companies (see related article here >>). It should be also abundantly clear to everyone that BI does not refer to reports browsing, stats perusing. It’s a bit more complex than that! Any travel business calls for BI, but its implementation will be fruitful as long as minimal requirements are met, and certain shortfalls tackled… 

To each, their own cup of tea

While prospecting for travel-tech among mid-small sized companies, it’s very easy to find two kinds of people: those holding budget but lacking expertise, and those lacking both. Neither would buy, because they usually search for alternatives that, they assume, would cost less than third party products.
Allow me to illustrate the evident sophism in this reasoning with an example. The marketplace offers dozens of free CMS options, and as many free widgets or plugins for reservation systems. Supposing you’re a small hotel or tour operator owner, if you do it yourself you would save between 300 and 1000 euros for a professionally looking website, sporting a booking engine. It would take about two full working days from set up to publishing… as long as you know how to do it. If you don’t, it would demand at least an online (free) course, one or two borrowed books, and paying some freelancer to fix the mess you’ll end up doing. Is the second option really cost-free or cheap? Add up the hours you’ll need for the course, reading the books and actual website building, and multiply that figure for your hourly wage. If the result is less than a thousand euros, let me know. I may have a job for you.
Now let’s expand the idea to a multinational bed bank or hotel chain, for instance. That kind of company typically has a medium to large IT department, populated by very capable engineers and developers. Yet, those IT teams are -by definition- very specialized, due to the nature of their daily tasks. Enter the enlightened top manager, who’s decided it’s time to put data to work. “Hey nerds, I want a BI system by next week!”, the savant would command from the grand stand. Sure enough, the computer wizards will deliver in a semester or two, as expected. Problem is, the resulting “BI” creation will -99% of the time- be useless or unused, quickly forgotten in spite of its sophistication. Again, many production hours wasted for no measurable result, in pursue of hypothetic money-savings.
Both cases, the SMB owner and the multinational corporation, show a blatant contravention of the 80/20 Law (Pareto Principle >>), not to mention a total disregard to the Lean methodology. Who cares, these are rather startup concepts, somebody would argue. Right, I don’t agree with that, but point taken.
Aside generic fallacies, here’s my list of endemically spawned reasons for BI project failures in travel businesses:

Miscommunication between IT and management: this is quite obvious. Each side speaks a different language, sees business under a different lens. One has an extremely practical approach, the other must rely a lot on creativity. One contemplates supply and demand in commercial terms, the other thinks it’s all about push or pull, bad or worst APIs. Even so, they have a common trait: none has got the time, patience or skills to deeply dig into BI instruments and methods.

Difficult marriage between data and internal processes: this is not a technical challenge, only a consequence of the previous issue. Managers know exactly what they want, but are unable to explain it in IT terms, while programmers are too busy (or single-minded) to realize that coding might not even be necessary for whatever the managers asked. To worsen things, adoption of third-party solutions is inevitably a trial & error process, if there’s no previous knowledge or the vital skills to use such solutions.

Top to bottom commissioning: the natural outcome of the first two problems. Picture the top manager from the earlier example… Just for fun, assume it’s a middle-aged guy. He’s been around, fought many battles, knows his stuff. He’s heard too many times the buzzwords, so he decides it’s worth trying at least one or two of them. Summoning the IT posse, the top man orders a clever and complete (from his point of view) project that would result in a collection of KPIs displayed on a dashboard… by spending the most ridiculous amount of company’s resources, both sides of the money gauge, depending on company’s size and manager’s power. Best case scenario, the result -after months of development- consists in a static console that could’ve been done in a bloody spreadsheet or any stupid freeware.

Internal politics: while the first three problems in this list are ubiquitous, internal resistance is a symptom commonly found in larger organizations. Some departments may oppose alien methods (or external leaders); IT teams could rise against (or even boycott) whatever they deem unproductive stuff. Let’s be honest here:

  • Non-technical personnel have a hard time adapting to anything that slightly deviates from routine, and usually there’s nobody available (or capable) to onboard them into new interfaces or procedures. I can relate!
  • Ninety nine out of a hundred computer scientists (especially senior ones) hate messing with business data, which is only a dirty byproduct of their brilliant coding. I second that!
Absence of a “Dragoman”: I’ll describe this one towards the end of the article, please bear with me.

Here’s a helping hand

You surely noticed a pattern there: the source problem is people, not technology. People that think they’re milking business intelligence because they read daily Excel reports. People that believe acquiring BI tool licenses, places their organization at the avant-garde of data processing, already ahead of competition.
For all my unapologetic rants, you should know by now that I loathe criticism for the sake of it, without at least suggesting simple remediesI compiled the list above based on personal observation. I’ve learn from my own and other’s mistakes, so trust me when I say:
[ctt template=”3″ link=”kx3mz” via=”yes” ]Business intelligence is NOT an IT function! It is a data-core survival ability. As such, either management and IT get the relevant competency, or they outsource the task to specialists.[/ctt]
Similarly, the technology should be bought, not made. Up to a decade ago, it made sense for large enterprises to code their own BI platform, ducking steep license fees. Mind, it’s not only a matter of building the system: once deployed, it will need maintenance, be flexible for future (or unseen) needs, get periodic updates… Exactly like any other complex piece of software. Nowadays, it’s preposterous to take the DYI approach. But wrong tech selection can be an even more expensive exercise, avoidable only if the required data skills were duly assimilated, or if consultancy was not neglected.
Once tech and skills were sorted out, I’d recommend to move in little steps first. No point in planning and executing an all-departments data-driven culture shift >> at once. No use for machine learning algorithms, if nobody pays attention to the company’s website logs first. Better to start with small projects, one department at a time. Otherwise, cultural shock will prevail, information will be condensed back in silos, despite the colorful dashboards implemented company-wide.
Above all, any kind or size of travel enterprise will be successful in home-made BI if the critical element is not missing: somebody balancing technical knowledge and domain expertise. In other words, the “translator” between the IT and business tribes. The guy or gal that knows integrations and databases, finds correlations or causalities, ask relevant business questions, and finally conveys the information to both sides of the fence. A modern Dragoman >> who eases communication between the white-collars and the keyboard warriors. The single person that can crush the first three problems of my list, just by being there!
Domain expertise on BI
Hence, companies pondering BI adoption should crunch numbers and perform research, not only technology-wise: Dragomans rarely are born and bred from internal ranks (except in OTAs), so outsourcing them as well is the smart option… Alas, they’re not easy to come by in any industry, for the same reasons I pointed out in the list above. That, ladies and gentlemen, is why mid to large sized travel companies consistently flop on their BI projects: they place as frontrunner either a business champion or an IT witchdoctor, who can (barely) coexist only as husband and wife, working in different verticals.
Simply put, it doesn’t matter at all which tech you favour, how large your IT staff is, how fat your wallet is: no Dragoman, no BI. Become one, raise one, pay one… or keep getting “intelligence” from your beloved tabular reports.
Thanks for reading and sharing!
Marcello Bresin