The agony of travel search and the eyeglass shop paradox

With the proposed acquisition of ITA Software by Google, Kayak‘s apparent serial successes, Hipmunk buzz, Everbread expectations, and many more, it looks like the search problem is still looking for the best answer.


But with “best” being the mother of all ambiguous words, one approach is to measure search quality by technical parameters.

This can includes things like number of sources searched or number of pages/fares searched; speed to respond and sped to respond with lots of results; freshness of data; accuracy of results; amount of details returned for each option; relevancy to the query, and many others.

There are companies who specialize in comparing search engines, they run lots of parallel queries and measure who is able to find the lowest price. Inevitably, there is no absolute winner, results are provided as percentage of times where one search tool beats the other.

But is lowest fare the “best” measurement criteria?

Fifteen years ago, the market was delighted with innovative tools from the GDSs to automatically find the three best fares, then the nine best fares, then best fares also including two connections, then 200 best fares, etc, etc… it was all about best fares.

Today we seem to be back to a quest for quality, with tools trying to avoid flooding users with hundreds of options and suggesting the best options (cheapest fare, lowest CO2 emission, lowest travel time…).

But I question if either approach is better than the other.

Focus on the user need

A “good” search tool is one that helps the user find what they are looking for, and this is where it gets interesting.

Sticking with travel, in some cases the traveler is looking for the best price, in other cases for the best schedule; in some cases for the preferred airline, in other cases for the shortest connection; in some cases for a specific airport, in other cases for any airport. Parallel criteria apply to selecting hotels. And even for the same individual, the decision is different at any new trip.

The innovation brought by metasearch players has been to provide access to a very wide range of options, and let users manipulate filters across different dimensions to slice and dice the results set.

The search engine role is actually just to return as fast as possible a large dataset – the user will use different filtering criteria at each trip according to the priorities of the day.

But, most importantly, the user can make trade-off decisions after having checked the range of available options. This is not a trivial statement: enter eyewear.

Lessons from an eyewear shop

Someone I know comes from a family who’s been running an eyewear shop for 40 years.

The moment a new customer steps into the shop, the owner can tell exactly which glasses they will end up buying. But he also knows something else: that he cannot tell them.

The customer will have to spend hours looking at and trying on all different models. Only at this point the customer will feel in a position to make a decision, and to buy the glasses.

The owner could save customers precious time, tell them upfront what they need and close the sale in 60 seconds. But if he did so, the customer would not be in a position to take a decision – and would go somewhere else.

The insight: a purchase decision is a function of the options available.

With this in mind, we can understand a number of things:

First, the roots of online travel: instead of just getting one or two itineraries from an agent, the web made it possible to directly check “all” options from the comfort of our desk and feel like we’ve taken the best decision by accessing all information directly.

We still do so today – feeling compelled to check across a multitude of sites to make sure we’ve explored all options. But are there too many options? Maybe…

Enter metasearch. Don’t waste time checking out all those sites, they’ll do it for you – and they’ll give you the tools to easily compare, slice and dice according to your needs.

We know in fact there are limitations in the reach of each search engine, but if users trust the brand, it’s enough. Give users the sense that they are taking an informed decision.

What’s next – agony filters?

I believe there is still plenty of innovation possible in travel search, but success will come from the user experience. Give the user the feeling that they are taking the best decision.

Give a sense of the full picture, empower them with tools to check all details and, of course, keep it simple. But which travel search engine makes you feel like you don’t need to check anywhere else?

Share on FacebookTweet about this on TwitterShare on LinkedInEmail to someone
Daniele Beccari

About the Writer :: Daniele Beccari

Daniele Beccari is a contributor to tnooz, and head of travel products at Criteo.

As travel technology strategist, he has helped startups and blue-chip corporations define and launch innovative solutions in leisure, corporate, online and mobile sectors. He also served as Vice President, Europe and B2B, at Isango! (now part of TUI), and previously as head of corporate products for the e-travel division of Amadeus.

He started his career at HP, working on what is known today as the Internet of things. An MBA graduate from INSEAD, Daniele can be found somewhere between Paris, London, Turin, San Francisco or Tokyo.

Daniele's views are his alone and not the views of his clients or employers.



Your email address will not be published. Required fields are marked *

  1. Michael Raybman

    Daniele, this is a great article, hits home with what we do at WaySavvy. I think the big challenge for improving the search experience is three-fold: data integrity, user experience, and the actual search algorithm. I believe there is value in doing the busywork for the user and presenting the 3 best rates, as long as the rationale is exposed and alternatives are available and easy to compare.

    I would argue the problem in the travel space goes a bit deeper than in an eyeware shop. In the shop, you will probably not find two identical pairs of glasses on different counters. In addition, certain services in the travel market are objectively worse than others for most people. For example, most people won’t choose a flight that’s more expensive and has more layovers than a cheaper, shorter one in the same time slot. (Of course, there are always exceptions, some people might try to optimize for frequent flier miles, etc… but I’m talking about the average user).

  2. Richard Moore

    What a fascinating article! Actually the last paragraph pretty much sums up what I have been trying to achieve over the last couple of years. For that very reason I have kept the sources of our site really dry (FCO, NHS, embassies, etc.) so that people believe in the integrity and independence of the information on the site (not seeing it as commercial spin), and can therefore put their trust in it.

    I’ve always believed that if people do find a site that provides complete unbiased results for them, and airlines etc. guarantee the lowest price on their site, there really is no reason to go anywhere else. That way consumers search and make the decision themselves, as per the eyewear analogy.

  3. Martino Matijevic

    Bravo Daniele, excellent article and brings up some interesting questions. Your analogy with eyeware shop can be transposed to the role of a travel agent: he/she will guess the correct destination/budget/preferences within minutes of chatting to the customer, something metaserchers will struggle to do, however many ‘agony aunts’ filters you add.

    I love the question: But is lowest fare the “best” measurement criteria?” As you said, “in some cases the traveler is looking for [..] the preferred airline”. Indeed, only just yesterday, we launched the airline rating system, which shows on a scale of 1-5 the airlines who they perceive to be performing best.

  4. Martin Collings

    Great analogy with the eyewear shop. I like Aaron’s point about people not always knowing what they are looking for – I predict the linear nature of the booking path today being replaced with offers tied to real availability based on prior user behavior and being displayed on third party sites that result in a click through directly to the passenger details page or the purchase page on an airline or OTA website. Personalization meets the future of travel search/find. Unfortunately I don’t have a timeline for when it will happen or who will be first.

  5. Timothy O'Neil-Dunne

    Let me add if I may a little contribution to this debate. In the world of search we tend to think in conventional products about a SKU approach. There is a simple binary thing. Its there or its not. But Travel products are incredibly complex. Not only am I different from you but heck I am even different from myself. I don’t behave consistently.

    Separating the issue of the product itself which is a relatively simple concept – if we take for example an airline ticket from the attributes of the “what the heck do I do with the product” creates a number of subjective and not necessarily rational decision elements.

    Conventional logic deployed in Search today would tend to favour the commodity approach of the lowest fare. As Michael O’Leary is often quoted. That an airline trip is just like a bus. Its a utility.

    But the truth is that Travel is a highly emotive as a subject. We are also highly subjective in our approach to it. Travel is not in of itself a utility except to all but the most jaded of frequent flyers.

    What I think the airlines are now recognizing is that value is darn hard to demonstrate. The old days of a brand cachet doesn’t necessarily keep its relevance in today’s “Social” world. However as they have seen the ability to de-commoditize the product is now emerging. Making that work with Search is no small matter when Search has been so gamed by so many different forces.

    Ultimately the fundementals of how Search works conflicts with the infrastructure of the stuff at the back. I have written about this frequently before. We actually stand at a cross roads where the process of SEARCH can be replaced by a process of FIND. That experience can be achieved by direct and indirect players alike. In my view this is where we will see some effort and some key players work hard to address that very conundrum. I am really quite excited to think that someone in a back room somewhere has that problem on a workbench and is just whittling away at it.


    • Daniele Beccari

      I like this. I’ve always wondered when someone will actually come up with a “find engine” instead of a “search engine”.

      My feeling is that “finding” is a pure psychological dimension which is parallel to the actual technical marvel within the engine.

    • Aaron Ritoper

      Good point that people wish to find not search. I’d go one step further and suggest that people want a recommendation because in fact they may not even know what they are supposed to be looking for. Smart technical platforms are getting better, and I think there is still room for strong, well defined promotions and editorial content.

  6. Aaron Ritoper

    Good article, Daniele. Having worked in flight meta search for some time now, I have a few thoughts on what makes a good flight search website. Meaningful evaluation criteria deserves more attention from those within the industry and consumers. My point of view is that the biggest “agony” related to flight search is bad and misleading data. While many sites scramble to add new features and gadgets, I like to see sites that focus on the fundamentals. Flight search should have a very clear purpose. People deserve to see accurate prices, fast results, and all the best (available) market prices. We measure these factors at and it’s rather surprising to see how poorly popular websites perform on the basics. Sorry for the boring comment. I know speed, price, and accuracy is not very flashy. My opinion is that these simple considerations should be on the “what’s next” list.

  7. Michael Smith

    To partly para phrase Donald Rumsfeld, the former US Secretary of Defense, we don’t know what we don’t know. This is especially true for customers that are making a trip/journey for the first time. They might not have a clue where to start.

    The advantage the eye shop owner has (again, pardon the pun) is that he can see the customer. In this case, those visual clues provide him/her with the information to work out what the customer will buy based on experience.

    So, that begs the question, how do you know your customer? So, will the metasearch people ask customers (could those be your agony filters?) what their preferences are? Which could lead to the follow up question, how does the customer know what their preferences are?

    SkyScanner, I believe, has a tool which provides some inspiration on destinations which could, perhaps, be extended to do something similar for these options. That can then start to help give information and help people through the purchasing decision.

    It is a fascinating area and one which I see (again, pardon the pun) very little happening from a pure airline side of things to help solve. People like American have an option to search by fare or schedule, which is a start, but it is still driven by Origin, Destination and Dates (and, yes, some people have a fare matrix for dates and so on).

    Will be interesting to see if Google and ITA solve this issue for customers as well as charging more for the targetted ads to airlines and others in the travel market!

    • Daniele Beccari

      Mike – assume you know a lot about the user, and you could show the user just the 3 “best” options instead of 200. That’s I heard Everbread describing at ITB as part of their vision for search. Assume it’s a “perfect” result and the technology is really smart.

      Do you think that would be sufficient for the user to take a decision? See Stephen’s comment above.

  8. Stephen Joyce

    Excellent article Daniele. I think your analogy of the eyewear shop is quite relevant. It demonstrates the innate need for consumers to commit to the purchase by making the decision themselves. When they do this, they take ownership of the decision and are much less likely to change their minds. I’m not sure if a decision on airfare requires quite the same level of decision making however, but the premise is the same. Give the consumer the options, allow them to filter based on their requirements (whatever they may be) and let the most appropriate choice win. Decisions are often about compromise, it’s not always about price.

  9. Alex Bainbridge

    I like the google approach, you put two keywords in, get overwhelmed with results, and then start refining. Least then you know to start refining as you know the effort will be worth it…

    Overwhelm, hook the customer so they know you have the solution and then refine. Simple 😉

    • Daniele Beccari

      My question to you Alex is: when do you stop? What makes you feel confident that you have found the best possible answer and don’t need to search elsewhere?

  10. offexploring

    A good read to go along with this would be “The Paradox of Choice, why more is less” by Barry Schwartz ISBN -13 978-0-000569-6. Giving the user too much to choose from may mean they never do.


Newsletter Subscription

Please subscribe now to Tnooz’s FREE daily newsletter.

This lively package of news and information from Tnooz’s web site provides a convenient digest of what’s happening in technology that drives the global travel, tourism and hospitality market.

  • Cancel