Revenue Management

Revenue Management Louis Busuttil Consultant lbusuttil@hotmail.com Frequently asked questions about revenue management • • • • • • • • • • Why sell more seats than you have seats available for sale? Why charge so many different prices for the same product? Why do your prices have so many conditions attached? So what is revenue management? Is it the same thing as yield management? Why do you need computers to help with forecasting? Why do you need computers to help with optimisation? Are there real-world external barriers to implementation? Are there real world internal barriers to implementation? What‟s the current state-of-the-art in revenue management? Why sell more seats than you have seats available for sale? overbooking limit 300 C A P A C I T Y 365 Days prior to departure economy class booking curve 0 Why charge so many different prices for the same product? P R I C E inelastic economy class demand curve 100 elastic CAPACITY CAPACITY 350 200 150 100 CAPACITY price/demand relationship, market segmentation, rev = area under curve Why do your prices have so many conditions attached? • Anti-marketing: or how to make your product less attractive to the late-booking high-yield customer 350 – – – – stay Saturday night 200 150 book three months in advance 100 non refundable/endorseable/reroutable CAPACITY you can even get them to dress as Elvis! People will still shop around for the lowest price they can get, but you can force them to progressively “trade up” to the more expensive seats as you approach day of departure. So what is revenue management? Revenue mgt is the integrated control of capacity and price - pioneered by mega airlines who could fund the R&D costs, rev mgt systems can now be bought “off the shelf”, but systems-integration, data-quality, and business-processimprovement remain major challenges - other industries following (hotel, car rental, broadcasting…) - a must-have for high-fixed-cost, low-margin, pricesegmentable businesses “The airline business is won and lost at the margins” Don Burr “An airline applying rev mgt techniques head-to-head with one which doesn‟t, can enjoy a 10% rev advantage” Peter Belobaba Is it the same thing as yield management? 300 seats: full fare at $1,900; discounted at $1,300; would you prefer: a) 50 full fare and 250 discounted, or b) 190 full fare and 50 discounted, or c) 135 full fare and 135 discounted? The one that makes you most money (c) is not necessarily the one that gives you the highest average yield (b) or the highest load factor (a). Revenue a) (50*1900)+(250*1300)=$420,000 b) (190*1900)+(50*1300)=$426,000 c) (135*1900)+(135*1300)=$432,000 Yield 420000/(50+250)=$1,400 426000/(190+50)=$1,775 432000/(135+135)=$1,600 Load-factor ((50+250)/300)*100=100% ((190+50)/300)*100=80% ((135+135)/300*100=90% Revenue management processes need to be supported by systems to help with passenger demand forecasting and fare mix optimisation. Why do you need computers for demand forecasting? There are many human biases in forecasting: - treat more easily available / recallable data as more significant - confidence in uncertain estimates often higher than warranted - gamblers fallacy: assume that random processes „self-correct‟ - overemphasise conclusions from small samples: anecdotal evidence - distorted perceptions from order of data, scaling of graphs, summaries - start from solution and insufficiently correct for current situation - failure to regress to the mean, extreme values expected to continue - attach higher validity to info which confirms previously held beliefs - fact/value confusion: strongly held beliefs are presented as facts - success attributed to personal ability, failure to bad luck - order effects: eg undue importance on first and last items presented - seeking information to support views, wishful thinking - conservatism: failing to use new info to significantly revise estimates Why do you need computers for fare mix optimisation? Computer can‟t be beaten - there is a right answer! - this is a mathematical, number-crunching, optimisation issue point of indifference 100 70 50 1A 1B 1C 1D Y B H T protection - but it relies on good forecasts which in turn depend on good data - garbage in, garbage out Expected Marginal Seat Revenue = fare * probability of selling that fare Economic point of indifference: 70% of $100, or 100% of $70 Are there real-world, external barriers to implementation? Even if you get the algorithms and the systems and the internal business processes right, you can‟t do this without external market discipline, and that means stopping distributors from undermining your efforts at revenue management through abusive practices. eg 1) Class abuse Book in Y (still has seats) 350 200 150 100 But ticket in H (even though it‟s closed) YBH T eg 2) Name change abuse H class available now, but anticipating a clamp down on class abuse, fill in some dummy bookings today, then as day-of-departure approaches when a real customer materialises, change Michael Mouse to Mr Smith. Are there real-world, internal barriers to implementation? Sales don‟t buy in to discipline “we‟ll lose business to our competition” Lack of net-net revenue accounting / yield-management capability Poor fare structure / inadequate yield management data maintenance Hi variance fare bands leads to “inversions” (higher-class-lower value trumps lowerclass-higher-value) 350 200 150 100 What you thought was $350 is only $190 when you take into account free limo/hotel 350 200 150 100 YBH T YBH T What’s the current state of the art in revenue management? AMS LHR FRA PAR ZRH ROM IST BAH DXB BOM SEL FUK NGO TYO OSA TPE HKG YVR SFO LAX YTO NYC BKK MNL KUL PEN CEB CHB SIN JKTSUB CNS DPS JNB ADL BNE PER MEL SYD AKL O&D control: managing the trade-offs between high-yielding point-topoint (eg 3rd/4th freedom) versus low-yielding connecting-traffic (eg 6th freedom) to maximise net-net revenue in real-time on a networkwide basis.