Over the past five years, much has been written in the hotel industry press about attribute-based selling, or ABS. I first covered the topic here about two years ago. At that time, there were lots of reservation systems and booking engines supporting ancillary sales (covered in part two of that article here), but no commercial products that offered guests the ability to specify the attributes of the guest room that matter to them. For those who are unclear, I will cover the difference between attribute and ancillary sales in a moment.
While ancillary sales have become commonplace in the past few years, attribute sales are still very early in their evolution. However, there are now several products on the market that support attribute selling at least to a degree, and some newer core platforms (think reservations and property management) that are designed to support attribute selling natively. IHG introduced a modest version on its own website and app last summer, although the implementation to date (via its Amadeus central reservation platform) is modest. Most of the other larger hotel brands reportedly have ABS on their longer-term roadmaps, although I believe they are likely years away. Developing a new reservations platform and rolling it out takes time; it took Amadeus and IHG nearly six years from the first public announcement on ABS (on slide 10) to brand-wide rollout of even limited capabilities.
The emerging ABS products I have reviewed, while far from mature, have significantly advanced the industry’s understanding of ABS. It is therefore time to revisit this selling approach, its benefits, its challenges, and one major myth. There is a lot to cover, so I will do this in two parts. This week I will cover how attributes can play into the sales process with other elements; I will mention the vendors who are active in the ABS space; and will provide some specifics about how ABS actually works. There are multiple models that need explanation.
Normally I would cover the question “why consider ABS” early in an article like this, but there are enough misconceptions about ABS in the industry that I found I couldn’t do that without first addressing some of them. Many readers who have misconceptions will likely dismiss the potential benefits of ABS without ever looking at it. An understanding of the specific deployment options will make it clearer who can benefit, in what time frame, and by how much. Consequently, I will defer discussion of the benefits – the “why” – along with operational and yield management considerations, to Part 2 (scheduled to be published on April 28).
The Role of Attributes in the Sales Process
Fundamentally, the sale of any regular hotel guest room can be characterized by four dimensions (hostels and vacation rental properties are a little different; I’ll ignore them here). While terminology varies, most of the experts with whom I have spoken are in rough agreement on the following structure and definitions, which I will adopt. Many product marketers have conflated all two, three, or all four categories and called them “attributes,” but that is simply to enable them to appear to be on the ABS bandwagon without actually implementing true attributes.
- Attributes are inherent characteristics of the guest room such as size, location, layout, view, whether it has a balcony or kitchenette, and the like. Attributes can also include other characteristics that could in theory be changed but are normally not, such as bed type or presence of a mini-fridge or safe. Attributes can never be unbundled from the specific room they are a part of.
- Ancillaries are additional goods or services that may be provided to the guest, such as spa treatments, breakfast, parking, flowers, or a local tour. Ancillaries may be delivered to the guest room, elsewhere in the hotel, or off-property. Ancillaries can almost always be unbundled from a specific room booking, and are normally unaffected by a change to the assigned guest room or room type.
- Policies include deposit requirements, cancellation rules, minimum length of stay, and the like.
- Duration specifies the time period covered by the booking. Most commonly this will be from check-in time on one date to check-out time on a later date, but many hotels offer early check-in or late check-out for an additional fee or as an elite loyalty benefit. Some offer day-use rooms and a few even rent rooms by the hour or minute.
Each attribute, ancillary, policy, or change in duration may affect the rate (and of course the rates may also depend on the number of guests, season, day of week, audience, channel, or other factors). Packages are simply rates that are applied to some combination of attributes, ancillaries, policies, and duration. And every attribute and ancillary (and if needed, even a policy) can have an associated inventory. Inventory enables the hotel to control how many sales are offered at concessionary rates; it is also essential for ensuring delivery of anything that is available only in limited quantities.
Ancillaries, policies, and duration have been around for quite a while and are deployed at some level of maturity in most commercial and brand-specific booking systems. Attributes are the new kid on the block and will be my principal focus here. With still-limited but nevertheless notable real-life deployments, we can now look at what works in different situations, and help hotels who are interested, but that have not yet embarked on the ABS journey, to consider their options.
Companies to Watch, and Other Acknowledgements
I am indebted to several companies who shared their approach to ABS, including three with products in the market that use ABS for selling and/or upselling: Expect Me, GauVendi, and Roomdex. GauVendi offers both an ABS-enabled web booking engine and ABS upselling for existing bookings (including third-party); Expect Me and Roomdex support ABS only for upselling after the initial booking. Amadeus is known to have some ABS functionality (visible on the IHG website) but did not respond to a request for an interview. Not yet in the market, but also worth watching, is a new company called rationalAI, founded by Pierre Boettner. Even several competitors cited Boettner as one of the most knowledgeable people in the industry on ABS.
In my research, I also found several companies whose recent-generation hotel operations and distribution platforms were designed to support ABS, but who were waiting for a lead customer to help define how they should deploy it. This is important, because most legacy property management systems (PMSs) and central reservations systems (CRSs) are fundamentally incapable of managing true attribute-based selling. Older systems require add-ons (such as GauVendi, Expect Me, or Roomdex) to manage availability and pricing of multiple attributes.
Some newer enterprise platforms claim to have been built to support ABS; these include HotelKey, Mews, and Shiji Group. Oracle also has fundamentally restructured Opera Cloud to enable ABS, deployment of which is now in its 12-month roadmap. My conversations with each of these companies generally confirmed that their back-end design should be ABS capable, and some hotel groups may well benefit from engaging with them now. But the caveat is that these products are not yet proven, and that each of them may have made design decisions early on that may limit certain aspects of ABS functionality. Furthermore, little if any work has been done on the user experience or tuning of the sales process, which are also critical to success. These are the elements that hotel companies will most want to customize.
In preparing this article and its follow-up, I also got significant input and wisdom from George Roukas at Hudson Crossing and Klaus Kohlmayr from IDeaS, for which I am most appreciative.
How Does Attribute Based Selling Actually Work?
Now that we have actual commercial products to look at, it is becoming clearer how ABS can work in the selling process. I will separate this into direct sales and channel sales.
Surprisingly, even many experts told me “you cannot sell attributes through online travel agents (OTAs).” That is a fallacy. In fact, the industry has been selling attributes through channels since the earliest days of electronic distribution. You probably never thought about it that way because in most cases, the only attributes offered were room category and bedding arrangement, and they were usually bundled together (one room category and one bed type defined one room type that could be inventoried and sold). The majority of third-party bookings are sold based on this combination of attributes, such as a deluxe king or standard twin. The only common exception is for run-of-house bookings, which are basically for “any room,” meaning a room with no guaranteed attributes.
Historically, attributes were typically packaged into “room products” like “A1K” (superior room, one king bed) or “C2D” (standard room, two double beds). But OTAs do not care hotels define products; if a different set of attributes and ancillaries will sell better, the hotel just needs to be able to define it, map it to actual inventory in their own system, and push it to the OTA channels. Many of the ABS platforms discussed here can do just that (or are designed to do so, for those not yet released).
The limitation on legacy systems (such as most central reservations systems) is not that they cannot sell attributes, but rather that the only way they could inventory them was to make them into room types. This is very limiting because every room in the hotel must map to one and only one room type, meaning you cannot sell the same room as a “king bedded room”, as a “king bedded room with balcony” and also as a “king bedded room with kitchenette.” You could decide to sell it as a “king bed with balcony and kitchenette,” but might lose customers who place no value on one or two of those attributes – even though the room might otherwise go unsold.
Direct Sales Models. In this version of ABS, implemented primarily on hotel websites, the user does not necessarily have to start by picking a room type. Instead or in addition, they specify the attributes they want, such as city view, high floor, and king bed. They might also specify policies, such as whether the rate must be refundable, and ancillaries that may be critical (such as reserved parking). The website identifies matching or near-matching products, ranks them, prices them, and presents them.
Some of what the website presents may be “standard” products that just happen to fit the requested attributes, while others may be constructed on demand. Prices are calculated by formulas and may include some rudimentary yielding, for example some attributes might be more expensive on weekends or in certain seasons. Some ABS engines even allow the guest to pick a specific room (a room number is just one more attribute, it just happens to only apply to one room). One vendor who supported multiple sales models reported that they got the highest room rate from guests who booked by starting with attributes rather than room type, but there is not enough data to make this a generalized industry-level conclusion.
Another variation replaces traditional room types with “products” that are defined based on a specific combination of attributes, ancillaries, and policies. The underlying inventory model still uses attributes, but the base sale options look more like traditional room-types, just with richer descriptions. A hotel may define dozens or even hundreds of products based on combinations of attributes, ancillaries, policies, each designed to appeal to a particular market or submarket. These products may then be “matched” to the specific sales opportunity based on whatever is known or can be inferred about the guest: party size and ages, length of stay, season, day of week, country of origin, language, past guest history, booking channel, or other factors.
Best practices suggest that only a handful of these options should be presented to any prospective customer. Analytics can be used to determine which products should be displayed for a particular sales opportunity, and in which order; or simpler approaches can be used, such as one set of products for lone travelers, another for couples, a third for families. Some products may be quite basic and defined simply to grab attention (many guests will gravitate to products labeled “Most Popular Choice” or “Manager’s Special”); others may be designed to capitalize on a short-term need to push attributes of rooms that would otherwise remain unsold.
One hotel that implemented ABS found high demand on a particular weekend for rooms with a street view that they considered undesirable – until it realized that guests wanted those rooms because they offered a good view of a parade that weekend. In a case like this, the hotel could define a “Parade View Room,” perhaps package it with drinks and snacks, and sell it at a substantial premium – only for the nights around the parade. This kind of flexibility is a major advantage of the ABS model.
There were many comments from the vendors I spoke with, and even a bit of consensus, on tradeoffs between complexity and conversion. The premise is that the most important task on the booking page is to close the sale (any sale!), and that too many options or choices will reduce the conversion rate. Upsells from one of the offered products, by this logic, are better presented after the sale has been made. This suggests offering a few well-targeted products in each sale opportunity, based on combinations of attributes, ancillaries, and policies that are likely to appeal. Then in subsequent steps (in the same session or later on, any time up until arrival), a secondary page can suggest additional upsell options, whether attribute-based (“upgrade to high floor with view”), room-type based (“upgrade to junior suite”), ancillaries (“add breakfast”), or policy (“make this refundable”).
Finally, the direct sales model may be varied based on the customer type. For example, the ability to select a specific room in advance might be offered only to loyalty members who book direct, and might be priced based on loyalty tier. OTA bookers might be offered the same choice but only at check-in, with a subtle suggestion that that next time, if they book direct, they can select their room at the time of booking.
Indirect Sales Models. An ABS engine that allows the hotel to define products should be able to push them to most third-party channels and channel managers, just as with standard room types today. An OTA booker will not be able to start by specifying attributes without a room type, but can still select a product that has been designed with attribute, ancillary, and policy combinations to appeal to guests using that third-party channel. The hotel is not selling attributes per se in this model, but is still using attributes to create more appealing and better-targeted products, and to easily refine them over time.
To be sure, most OTAs today do not support shopping by room attribute (although the larger ones do support property selection based on property attributes). But you can still upsell room attributes or ancillaries to OTA customers. Most OTAs give you a way to communicate with the customer and (with some limits) will allow you to send links to web pages that allow the guest to customize their stay, explore the hotel, and maybe preregister. The landing page for the link can suggest upsell options that can include anything from adding specific attributes or ancillaries, to completely changing what they purchased. The incremental revenue booked on the hotel’s upsell page is usually noncommissionable.
To Be Continued
Stay tuned in two weeks for Part 2, scheduled for publication on April 28. That will cover the benefits of ABS selling, key operational considerations, and yield management of attributes.