Definitely Doug 1/21/22: Business Intelligence and Analytics: RIP Excel?

Hotel businesses are run in part based on numbers, and they need to see them quickly if they are to address problems early. But with the many different silos of information found in a typical hotel group’s technology portfolio, this has long been easier said than done.

Not very long ago, Microsoft Excel was the most commonly used tool for hotels that needed to combine data from multiple systems, whether within a single hotel or to facilitate roll-up reporting across a portfolio. Excel may still dominate today, but if so, it is rapidly losing share to business intelligence and analytics software. Despite this, there is a lot of confusion in the market around this category, largely driven by the hundreds (literally) of hotel software products claiming to provide business intelligence. Today’s column will try to clear that up.

If you’re buying a property management system (PMS), point-of-sale system (POS), channel manager, accounting system, or other hotel software packages, it’s not uncommon to find a module called Business Intelligence (BI) or something similar. They not only provide access to reports, but usually have the ability to drill down into summary numbers, filter based on dates, market segments, departments, or other criteria, and present information in graphical and dashboard formats.

But in my book, most of these do not really qualify as “Business” intelligence because they only work with data in a single, siloed system that falls far short of everything you need to run your “business” intelligently. To be sure, they can be hugely useful, but unless you are running your entire hotel business on that one system, they can give at best a limited picture of the overall business. Most of them focus on a specific function, such as finance, operations, or marketing and distribution; if they integrate any data at all from other systems, it is usually very limited and may require manual re-keying. And while a few (especially smaller) hotels may operate on a single system, most hotels of any size or complexity use multiple systems, often numbering into the dozens. A BI system that only has access to data from one of them will be very limited in its scope.

Not surprisingly, many of the BI systems in common use are built into financial reporting systems, because those systems have audited, cross-departmental data that plays a major role in performance measurement. They have their place, but also their limitations. Data are (or at least should be) correct, but are typically limited to revenues and expenses. You can see what happened, and can drill down to see detailed sub-accounts and even individual transactions. But you often cannot see why, because that detail may be buried in the PMS, POS, or labor management system.

Why were payroll costs so high in the food and beverage department yesterday? Your financial BI can probably tell you that overtime pay was responsible, but not the reason, which might be that delegates from a large in-house group took over the restaurant bar last night and spent tens of thousands of dollars, paid for by a combination of credit cards and individual guest room charges; staff were scrambled to help, and the bar was open hours later than usual. That kind of information is useful not only in assessing the departmental payroll variance, but also in determining the value of a group beyond what gets posted to the master folio. But fully understanding it may require information from the PMS, POS, sales and catering, and/or time-clock systems.

It's also common to see BI modules in systems designed to manage hotel marketing and distribution. These can be very helpful in targeting particular markets and managing distribution costs. But aside from room revenue and distribution costs, they usually provide little insight into profitability. They typically cannot tell you what types of customers spend the most on food and beverage, spa, and other hotel services, because that data is locked in other silos.

By my definition, a true hotel BI system should have detailed data from multiple systems so that it can address questions like:

• What is my distribution cost by market segment?
• What are the costs of my brand affiliation by market segment and loyalty level?
• What is my labor cost per restaurant cover, per room cleaned, per function delegate catered?
• What geographic regions, length of stay, party size, or days of the week generate the most revenue and most profit? How does that vary by market segment?
• What is the profitability of food and beverage, retail, or spa outlets by day of week and time of day
• What food and beverage items are most popular among the 10% of most profitable guests?

These questions are more interesting for larger and more complex hotels; many smaller and lower-market properties will likely find this to be overkill. For them, the built-in analytics of their financial systems may be more than enough.

True BI, beyond simple financial reporting, used to be something that only larger and more luxurious hotels needed, but costs have come down and managers have become more data-hungry, so BI has been rapidly moving into the midmarket. It may never be useful for a 50-room limited service roadside property, but it is now being adopted by many midscale hotels, especially those with significant food and beverage operations.

There are surprisingly few vendors providing what I consider true BI in hospitality – platforms that integrate significant amounts of data from multiple systems into a single platform. I spoke with key executive at several of the leaders and innovators in the space, including Avenue9 Solutions, Datavision Technologies, inhovate, myDigitalOffice, and Profitsword, who provided much of the wisdom included in this article (for which I thank them!). I am not going to try to compare them because while they are all good products, which one is “best” will depend almost entirely on the fit with your needs. But looking at several of these (as well as any others that come onto your radar) will ensure that you can make a well informed decision.

In addition to products like this, there are numerous large consulting firms that will build custom (and expensive) solutions, usually for very large clients, and various do-it-yourself approaches using commercial tools and a team of data scientists. But today I will focus on commercial packages.

Key Considerations

Before evaluating BI products, it is important that you first define what you are trying to achieve. What are the key insights you are looking for, that you cannot get today? Audited financial data is necessary and can serve many purposes, but often is not actionable – you can see that something went wrong, but may get very little indication of why, and it is historical only (but no action you can take will change the history). Unaudited data can be more timely, more actionable in some cases, and more detailed, but also less correct. And owners or asset managers will care about different things than a general manager, marketing director, or food and beverage director.

Do you need only historical data, or forward-looking as well? Budgets, forecasts, business-on-the books, and revenue management system analyses can be combined to give a better look at the future, and overlaid with actual results to see the past, present and future together. This is a big differentiator of the systems I saw; the better ones know how to combine multiple sources, for example taking last year’s results from audited financials, last month’s from trial balances, yesterday’s unclosed results from the financial system, today’s and tomorrow’s mostly from on-the-books reservations, next month’s from now based on on-the-books plus a revenue management (or other) forecast, and next year’s based on the budget. This gets tricky when you start to drill down and need to change the source because (for example) the audited financials don’t have transactional folio detail – but some of the systems are quite good at handling this.

Another important question is how you want to see your data. Some people like visual presentations in charts and graphs, others prefer tabular presentations of numbers. Some hotels may be fine with canned reports provided by the vendor, others will want to modify them or create some of their own, while still others will want to access the raw data directly to see what they can learn. Most of the systems offer quite a bit of flexibility and the ability to maintain multiple presentation formats for different users; tools for working with raw data vary. A good piece of advice is not to simply recreate a report you already use in a new BI system. Rather, you should engage with the vendor to find out other ways the information might be presented. You might be surprised at how much more useful a presentation might be if can be if you approach it with an open mind!

A very useful feature is the ability to view the original data as it was imported from the source system. There will inevitably be data anomalies in the BI system, and this will help track down whether the issue existed in the source system or was introduced during the import or normalization. If the BI system is ingesting a report from the PMS, can you view the actual report from which a particular number displayed in the BI application was obtained?

Several of the vendors I spoke with counseled keeping your objectives modest, noting that most hotels do not know how they will actually use business intelligence until after they get it. So rather than designing a solution to do everything you can think of, it’s better to pick just a few high-priority things where you know you need better data and insights. Once you have those solved, identify another few objectives; then rinse and repeat.

Once you know your objectives, there are a number of questions. First, where is the data that is needed stored? Can the solution get access? If yes, does data use common definitions and a common format, or will massaging be needed to (for example) compare data across properties that use different PMS systems or market codings? Do you need to be able to access and analyze individual transactions (such as line items on a PMS folio or POS check) or just summary financial data, such as daily departmental revenues and costs? This exercise should result in a list of systems from which you will need to extract data, and the knowledge of how much detail you will need.

Armed with this list, the questions to ask your prospective vendor(s) are (a) what systems do they already support data extracts from; (b) for the ones they support, what is the level of detail? Some systems extract daily summaries for predefined groupings such as market segment, while others get every underlying transaction (reservation, guest folio, POS check, guest survey response, etc.) and compute the summaries themselves. If they do not ingest the detail, then you will be limited in the questions you can ask. You might be able to find out what happened within the system’s predefined groupings (such as market segments), but you will not be able to drill down further or redefine the groupings. Consider some of the sample questions below and if they are ones that matter to you, ask yourself where you store the data that will be needed to answer them.

This exercise will likely eliminate some of the prospective vendors. It may also help you refine your objectives, because it will give you a better sense of what can be done quickly and inexpensively, so that your BI effort can produce some quick wins.

It is not necessarily a deal-breaker if a vendor cannot already extract data from a system you need, but it is a good idea to consider their relevant experience. A vendor that has extracted and integrated data from 20 different spa systems already, will have a pretty good idea what it will take to add a 21st to meet your needs. On the other hand, a vendor that has never done a spa integration before will have a steep learning curve, adding both risk and (probably) cost to the project.

A third key question to ask if you have multiple properties is how the vendor normalizes data. Normalization enables cross-comparison of data among properties using different underlying systems, coding schemes, accounting schemes, and currencies, such as is frequently the case with management companies that operate hotels under multiple brands or in different countries. If your hotels have to comply with multiple sets of accounting standards and you need consistent financial comparisons, then normalizing to a common schema such as the Uniform System of Accounts for the Lodging Industry (USALI) can be very helpful, and some vendors can do this.

If you need to normalize market segmentation from the standards of several different hotel brands, some vendors have predefined mappings that may work for you, or you may have to decide for yourself how best define them. From what I saw, some of the products make this process simple, others can require a significant investment of time and/or money to set up.

Lastly, an important question is how data from different sources are connected. For example, if you want to be able to analyze profitability of specific group customers, you might need connections that attach reservations and folios (both master and guest) from the PMS as well as checks from the POS, spa, and golf system to a guest in your customer or loyalty database. The data may come from different systems that either do not know how to make the connection or can only do it in certain specific circumstances, so usually the connection has to be made by the BI platform.

Pricing is always a key question; you can spend as little as $100 a month for a good BI system for a single small hotel, or many thousands for a large complex one. The cost will vary depending on complexity, the amount of data, and other factors, which is another reason it’s smart to start small and get to know your vendor. One thing to check with the pricing model is whether additional users incur an additional license fee (and if so, how much). BI works best if all of the managers in your organization can use it, so you will want a plan that will be affordable when they all do.

What Systems can be Connected?

I asked the vendors I spoke with what types of systems they ingest data from, and accumulated a long list. To be sure, no one vendor handled everything on the list, the younger vendors had somewhat more limited lists – but ones that could still meet the needs of many hotels. The list of potential systems includes PMS, financial/ERP systems, POS, sales and catering, central reservations, channel managers, golf, spa, retail, guest satisfaction surveys, social media reviews, competitive data such as STAR reports, weather data, exchange rates, airline schedules, economic data, forward-looking data such as from OTA Insight or revenue management systems, Google Analytics, e-Procurement, time and attendance, and labor management. On future product roadmaps but not available in the products I looked at today, were data from housekeeping and building management systems. The list will undoubtedly continue to grow.

As important as which systems are connected is what detail they can provide. In today’s labor market, some hotels want to look in detail at housekeeping productivity to understand what types of customers have low vs. high impact on housekeeping labor. And while a BI system may already get payroll from the financial system and hours from the time and attendance system, that provides no insight as to how long the housekeeper spent cleaning a particular guest’s room. That data is tracked by some housekeeping systems. If you can’t get it, either because your housekeeping system does not track it or because the data can’t be handed off to the BI system, then your BO solution will offer no insights on this question.

What Questions Can a Good BI Platform Answer?

Hotels who have never used BI are often at a loss to understand how they might use it. I asked about the kinds of questions BI customers wanted their vendors to be able to answer. To be sure, not every product can handle every question. But very few of the purely or primarily financial BI systems can handle many of them, so they are a good differentiator of what you get by upgrading to a true BI system.

• What determines the revenue and profit the hotel can earn from a particular type of guest? This helps target marketing, sales, and promotional efforts. Factors often analyzed include booking lead time, length of stay, market segment, size of party, day of week, seasonality, booking channel, nationality or geographic origin, rate plan/package/special, and even (for restaurants, spa, golf, etc.) time of day.
• Who are my top individual guests and what do they buy? What are the most frequently purchased menu items, wines, or spa treatments by the top 5% of guests?
• How much revenue and profit did a group provide, including individual guest charges that never showed up on the group master folio?
• What guests with future room bookings have used hotel facilities (spa, golf, dining, etc.) on past stays, and can we reach out to them before their stay to ensure their needs are met and our revenue maximized? Are we adequately stocked with menu items or wines that high-value guests have purchased on past visits?
• Which spa therapists are most requested by top customers?
• Which food and beverage servers are most productive (meaning more revenue per minute worked) or earn the highest tips per dollar sold?
• What are the distribution costs and brand fees (royalty, loyalty, etc.) associated with specific guests, market segments, seasons, etc.

It’s important to understand that once you have a BI, there is virtually no limit to the questions you can ask at no additional cost, based on the data that are loaded and the capabilities of the platform. You can build your own reports and graphs, you can filter data based on your own criteria, you can change those decisions in an instant to do “what if” analysis, and you can for the most part present it the way you want it. You will typically face additional costs only if you need to ingest new data or modify how it is processed as it is brought in. Yes, there are tasks where you may need paid vendor support, but if you have selected the package well, they will be few and far between.

Other Considerations in Evaluating a BI Solution

BI systems have lots of data, and data can be used for lots of purposes, some of them extending beyond what would normally be called analytics but nevertheless often providing real value to management. Some of the interesting applications and capabilities I ran across in my research include:

• Using a BI extract to prepare daily closeout data for a POS to the financial system to eliminate a multi-hour daily, manual, error-prone process.
• Combining various data sources and analytics to produce a detailed forecast of future activity for a hotel spa, combining reservations on the books, predicted bookings from transient guests with future room reservations in the PMS, and estimated walk-ins. Since most spa bookings are made very close in, this gave the spa manager much better visibility of future demand and scheduling needs.
• Using data on what facilities guests actually used in the hotel to customize post-stay guest surveys and ask only those questions that are relevant for that stay.
• Tracking of identified problem areas, where a problem is not only identified, but actions assigned, and follow-up scheduled to assess whether those actions worked.
• The ability to embed customized reports, even to the point of producing a “board deck” in PowerPoint with custom commentary and fully interactive data tables.
• The ability to save specific views of data and share them with co-workers with comments can be very useful for ensuring that insights are seen by the right people and acted upon.

Conclusion

BI tools have made great advances in recent years. With today’s minimal cost associated with storing even huge quantities of data, it is more and more feasible to implement a useful BI system that can provide very deep insight, and modest cost. Excel may be very cheap to license, but many hotels have found that BI systems can vastly reduce the amount of time needed to maintain even modest amounts of data vs. Excel. If you have not looked at the BI options in recent years, it may be time to do so!

‍