Revenue Management System: In-house or Outsourced?
Revenue management (RM) systems are ubiquitous in the travel industry, from airlines to hotels, rental car companies to cruise lines. The question facing many of these companies these days is whether to build their own in-house RM system or select a vendor-provided, outsourced solution. When Continental and United merged in 2010, both teams got together to decide whether to keep Continental’s PROS solution, or stick with United’s own homegrown Orion solution. In the end, and after very careful consideration, they chose to focus on their existing in-house solution. I know from numerous conversations with the team that it was not an easy decision. By all accounts, though, the airline has been very happy with since — admittedly after some hiccups with Orion and a successful transition to Gemini (c.f. Scott Kirby’s comments during United’s Q3, 2016 investor presentation, slide 42 and onwards). If you are considering replacing or upgrading your current RM system, I hope this post will help you make a wise decision.
A highly fragmented and competitive business
A quick survey of the revenue management systems (RMS) used by the top companies in air transportation (airlines), hospitality (hotels), car rental and cruise lines finds an even mix of in-house RMS and outsourced vendor RMS. Table 1 below shows that the largest airlines are split almost evenly between these two solutions, with PROS the most widely used among outsourced systems. The same mix applies, but with different vendors, across top hotels, cruise lines and car rental companies. In general, larger companies tend to be more likely to invest in their own RMS, while smaller players lean toward outsourced solutions. When expanding the above analysis to the top 250 airlines worldwide, in-house solutions are used by about 50 percent of the larger airlines and by 10–15% of the industry overall.
Key decision points
In my experience, the decision to go in-house or outsourced hinges on three major decision points: speed, flexibility and cost, and control. I will save for another time a discussion about two other important aspects of revenue management: governance and internal organization. These two aspects both influence and are a byproduct of the system decision you make. Drush and Horowitz have a good piece that discusses some of these aspects in more detail.
Let’s examine each of the three major decision points in more detail.
Speed is critical when implementing a revenue management solution. You want to realize the revenue benefits of your solution as quickly as possible, particularly in highly competitive environments where an effective RMS solution can mean the difference between profit and loss.
From the perspective of speed, the outsourced solution can look very attractive, given that the solution already exists and “only” needs to be implemented into the customer’s environment. While this process may sound relatively straightforward, it is usually anything but. In the airline industry, systems (e.g. distribution, reservations, scheduling, etc.) have to be identical or very similar around the world to allow for a relatively seamless travel experience. This greatly simplifies the process for RMS vendors, in that the majority of the systems they need to connect to are very standardized. Having said that, anyone at PROS or any other RMS-solution provider would emphasize that, even with such standardization, implementing an RMS requires a great deal of skill acquired only after years of experience.
Even when a vendor possesses those skills, implementing an RMS in an industry with standardized systems requires much more than simply plugging in a machine and flipping the “on” switch. In industries where systems may not be quite as standardized, deploying an RMS can be a lot trickier.
It may seem counterintuitive, but when speed is the issue, choosing to develop your own solution could result in a shorter time to value. If you design your roadmap to prioritize the development of those modules that you know will bring the biggest value — and design the system such that the modules can operate independently of one another — you could potentially realize the revenue benefits of your in-house solution very quickly, and accelerate the time to a profitable, full-blown, self-funded RMS.
Flexibility & cost
Next comes the flexibility and cost aspect of the equation. Specifically, is your organization flexible enough to redistribute internal resources to focus on this project? How much time can you commit to such an endeavor? And how much are you willing to invest in the project? Once again, the perceived advantage of outsourced solutions is that they should come online faster — at least in a quasi-completed form.
The challenge with in-house solutions is that they require a huge commitment of time, resources and budget from the company. They also require a willingness to weather the difficult times that invariably come with implementing complex, company-wide solutions. Building software from scratch is not easy, and progress can sometimes be agonizingly slow. Just ask the FBI, which contracted with a third party to upgrade its IT systems. The project failed rather spectacularly — and at a very high cost.
Vendor-provided solutions are usually much more cost effective — in the short term. Over the long run, however, the license and subscription costs that come with an outsourced solution can catch up to and eventually overtake the admittedly large upfront investment cost of developing a solution in house.
The question becomes whether you are willing to spend the money upfront. If so, have you calculated how long it will take to recover the cost of implementation? Surprisingly, anecdotal evidence I’ve gathered from conversations with colleagues who have implemented in-house RMS solutions suggests that, for larger entities at least, implementation and development costs are recovered very quickly — often within 3–6 months of the system going online!
Keep in mind that the spending doesn’t end once an in-house solution has been successfully implemented. Once the system is in place, you must provide on-going support and development or your system will slowly become outdated and an impediment to your business. Vendors have a strong incentive to stay at the forefront of technology. Companies with in-house systems all too often tend to follow the “if it’s not broken why fix it” philosophy. This approach may be cheap in the short run, but ruinous over time.
The final and perhaps most important decision point is control, which is better understood as a combination of: (1) science, algorithms & data, (2) roadmap and features, (3) confidentiality, and (4) security. While there are additional controls involved when building complex systems, these are the top four to consider, in my opinion. (Comments welcomed and encouraged).
Science, algorithms & data
The most obvious area where in-house RMS has a distinct advantage is with respect to control of the backend algorithms. While it is true in part that basic RM techniques are well documented, managing the specific quirks of your business will benefit from the use of bespoke algorithms or tools. Consider fare rules that apply to some specific markets, but not to others. Each market might require a very different approach to optimization and forecasting. Generic systems may be unable to support such an approach, while a custom system can be designed to do exactly that.
Another aspect of the science side is data. Despite the fact that data is of utmost importance in this day and age, numerous industries and companies are hard pressed to even gain access to their own data. Implementing an outsourced solution will require much more control than that. Anything from minor to major adjustments may need to be made to data sources to make them fit the required input formats. Creating this level of control can be both a good and a bad thing, but it is unavoidable when implementing a vendor solution. Choosing to develop an in-house system can provide more flexibility up front, enabling you to design the system to take full advantage of your data, even if that means that, in the end, not all that data you’re gained control of is used.
Another issue has to do with the output of the data into control systems. Revenue management systems need to distribute their optimized inventory and pricing into reservation systems and global distribution systems. This requires interfaces to these external systems so that both the input and the output data remain in sync. While these interfaces and connectors are usually available and developed by both CRS/GDS vendors and RMS vendors, these data exchanges can create friction between competitors. Imagine using an Amadeus RMS with a Sabre CRS: Sabre may not have much of an incentive to facilitate integration of the Amadeus system, since Sabre itself offers an RMS solution. This friction can arise with in-house solutions as well, although the CRS providers may be more inclined to work with the customer since there is no intermediary RMS provider involved.
Roadmap and features
Another very important aspect of control is the ability to create your own roadmap and feature development. Developing in-house means you aren’t competing with other customers to convince the vendor which features to develop first and to save for later. Developing in-house essentially means you gain total control over your own destiny when it comes to designing and building the overall system.
At the same time, working in your own “bubble” means you are not part of a community of practitioners, learning from your peers and adopting best practices. Vendors typically work with many clients with different needs and unique problems. This enables them to learn from their customers and bring these insights to their development work. If you’re isolated from this process, your solution may not include these best practices.
Confidentiality is a very important aspect of control in a highly competitive environment. I have often heard people point to the ability to keep a solution confidential as one of the positive differentiators of an in-house system. Numerous research findings, however, including MIT’s PODS consortium studies, have a different take on the issue. Research generally supports the idea that there may be a first mover advantage in RMS, in the sense that if one company develops a highly productive system it will have a leg up on the competition. But it also suggests that there is a clear global benefit to the industry when all competitors are operating at the same approximate level of sophistication. Contrary to the idea of a zero-sum game, companies — and, in fact, an entire industry — can actually benefit by sharing, at a high level, aspects of the RMS they have developed and found to work well.
In my mind, the jury is still out on whether or not keeping your secret sauce secret is truly helping. The Nash equilibrium suggests that when all participants use revenue management with the same approximate level of sophistication, everyone wins.
Data security is of paramount importance, and is a subject that often comes up when considering external software providers. It is a topic that I feel is slightly secondary to the previous three major considerations we’ve just discussed, but it is top of mind in today’s cloud-based world. Given the nature of revenue management systems and their necessary openness to a large number of external distribution systems, I feel that regardless of whether you choose an internal or external RMS solution, data security should be a critical part of your decision process.
It may seem that hosting your own system with internal hardware and data centers would be the more secure approach. However, in this day and age of cloud computing, an in-house system built with purely internal hardware is unlikely to be the economical choice. From there, you have to weigh the pros and cons of choosing a cloud-based platform for your internal systems or of relying on the vendor if you choose an external solution. Do you feel like your in-house team has sufficient expertise regarding data security? Will you need to hire an external provider to test your systems? Are you more comfortable with a vendor system that has been externally certified against intrusions? You must consider all these questions when making a selection, particularly as more vendors choose to embrace the benefits of cloud-computing platforms.
In my experience, I have found that the choice between in-house and outsourced RMS is never an easy one. Having been on both sides (the airline side at Continental and the vendor side at PROS), I have felt the pain of both customers and vendors. The decision truly boils down to three points:
How quickly can you get a system off the ground? How important is speed to your business? Can you afford to spend a few months to a few years to build an in-house solution, or will an off-the-shelf solution serve you better?
2. Flexibility and Cost
Do you have the financial means and resources to invest in your own in-house system and to support it for years to come? Or would it make more sense to essentially “rent” a system from a vendor?
Do you want to have the kind of total control over the solution, the roadmap and the backend you can achieve with an entirely in-house solution? Or would you benefit more from a solution informed by the best practices of an entire community of users, as you would be with an outsourced solution? Does your organization possess the internal resources to develop an in-house solution — and then support it over time? Are you ready to embrace the cloud? If you are, are you confident that your internal teams can prevent all external intrusions? Or it is better for your company to leave that to “the experts” — vendors that are in business to do that exact thing?
Once you make the final decision on your RMS solution, it will then become extremely important to confirm commitment from leadership and from the company as a whole. Implementing RMS, whether in-house or outsourced, takes time and hard work. Changing direction midway through the process can be difficult and costly.
You may have noticed that in the course of this analysis I did not discuss the people or processes associated with RMS. These are both, of course, very important aspects of the decision process. They too will have an impact on your decision. Consider how the revenue management function would be handled within your organization. If you develop in-house, is leadership aligned? If you outsource, does your vision match that of your vendor? Can you reconcile any differences, and possibly learn from a vendor solution? Drush and Horowitz, as referenced earlier, also provide some valuable insights into these aspects in their post.
Before I close, I should mention that there is actually a third choice: outsourcing the RMS function in its entirety. This is a very rare solution — one that few companies employ. Hotel groups sometimes use this approach with franchised properties, providing their franchisees with price and inventory recommendations through their central RM solutions. One can argue that these are not fully outsourced RMS functions, but they are as close as it gets today. I am waiting for the day when a revenue management vendor considers this solution as a possible end-to-end service for their customers. Until then, you’ve got a binary choice ahead of you: in-house or outsourced.
What’s your take on this thorny decision?