Travel Analytics Blog
Are Airlines Missing a Trick with their APIs
Posted by Sonja Woodman on Tuesday, September 18, 2018
Airlines are embracing APIs to reduce distribution costs. It means travel distributors such as aggregators and metasearch engines can get direct access to inventory and reservation systems to search for availability and facilitate the booking process.Suppliers benefit from a broader marketplace. Airlines have unquestionable expertise in brand.com analytics, but even LCCs, masters of driving down costs, are missing a trick by not analysing their APIs for greater visibility into their indirect channels.
The world of distribution, even among airlines, is increasingly embracing APIs as a way of reaching the indirect channel. Distribution via aggregators and metasearch engines will continue to evolve alongside the more traditional GDS, who are also evolving their services and connectivity.
Low Cost Carriers (LCC) traditionally focused on selling their seats via direct bookings, but many have forged distributions arrangements with the major GDSs. The impetus was to grow visibility and bookings from the less price sensitive business sector who still largely book through their travel management processes.
LCCs led the way in mastering direct channel web analytics, and most airlines have followed suit, keen to model LCC direct business success and cut out the expensive distribution channels such as the GDSs. But the business channel, especially those involved in long haul and multiple or mixed legs on a trip continue to rely on Travel Management Companies for their bookings, a trend that isn't going away.
It is possible to analyse indirect API traffic in a similar ‘google analytics’ way, but many airlines including LCCs still operate in the dark. While it is important to monitor API performance in real-time – to avoid unnecessary downtime or revenue leakage due to errors and poor response times, it is also possible and indeed compelling to analyse the business value of API content. APIs are the plumbing pipes that connect suppliers and distributors. In simple terms an API is a way for one company to access the products of another. APIs are basically the web services that connect and bridge the information flow across the connected supply chain – ultimately enabling the variety of options to be put before the consumer.
Today, searches for flights and offer replies are increasingly being exchanged using XML via a travel API. The NDC is one such example that is taking off. Open API can widen airline revenues by more readily making their content available to application developers and other potential partners. It can also lower the costs of bringing new content-using applications to market. Metasearch such as skyscanner is another important example of where APIs are actively in use.
Metasearch is here to stay and all of the airlines out there will have been forced to acknowledge the presence of their competitors on major sites like Skyscanner, Kayak.com, Fly.com etc. It’s really been a case of an ‘if you can’t beat them join them’ result. In the travel conference circles, there is an ongoing debate as to whether metasearch related bookings should be considered as direct or indirect business. Metasearch sites don’t work for free so there is always a fee to be made somewhere. Then there is the debate over whether CPA is better than CPC but we can leave that for another day.
At the technology level there are also different approaches and I am making the hopefully correct assumption that these no longer involve any form of screen scraping. In 99% of the cases the airline, whether that be LCC or Full Service, will have developed a specific metasearch API/platform that contracted metasearch partners are able to access. The airlines’ metasearch API provides flight price and availability information directly in a highly usable XML data format to feed the list of search results that we see appearing as we watch our flight options list fill in appear before our eyes often under a whirring wait banner.
Successful conversions appear in form a booking which has traditionally been handled by the metasearch API providing a ‘deep link’ directly back into the airline’s brand.com website booking page so that the booking can then follow the normal execution pathway. In more recent developments, the metasearch may take the booking directly (c.f. an OTA) or as recently announced by BA and Skyscanner, through an IATA NDC based API. Whilst both the metasearch site and the airline involved know all about the bookings – not surprising since there is the question of fees/commission – but how much does the airline know about the searches?
What's the Trick?
For any channel, the trick is to optimise the conversion rates. That means access to relevant analytics. While it is highly likely that most if not all of airline sales and distribution departments will have excellent website data collection tools and a digital marketing team to analyse every interaction to eke the last conversion out of the shopping flow, this is of little benefit to the API channel. In the API world there are no web page designs to consider only prices and details on rooms, rates and services to feedback as cold blocks of data. The trick is how can the airline using an API channel optimise the conversion rates?
Analysing and aggregating the XML message content enables airlines to get a clear picture of what products are being requested, what the responding offers are and how well they match. From this big data analysis, important nuggets of insight can be gleaned such as what category of travellers are buying what kind of flight experience, to which destinations and at what price. Detailed insight helps yield manage the price for revenue and deploy offers to satisfy customers. Data also reveals which suppliers are delivering real value in terms of bookings and which ones are adding to Look-to-Book infrastructure costs with little return.
What we are really talking about here is using predictive analytics to set on-line booking platform rules that will provide the real-time segmentation of airline searches to respond with the best available and most relevant products. Detailed data is required to feed the analytics and detailed data to measure the outcome of the rules which needs to be quickly and appropriately processed. Done correctly, this analysis, set rules and test sequence forms a cyclic feedback loop whose objective is continuous improvement. In the near future, it also likely that the slow manual rules setting process of today will be replaced by the rapid decision making AI technology that is already available.
At Triometric, we see the formulating and testing of the differential offer creation process as the perfect end game to a set of conversion optimising approaches that can be taken.
This is the first blog in a three part discussion. The next two blogs will talk about the Four Pillars of conversion that form manageable steps in this journey. We will also talk about the type of data needed and how it can be collected and processed.