Archive | April, 2008

TripIt and Dopplr – A Match (which could be) Made in Heaven

17 Apr

TripItI was recently introduced to TripIt, a “next generation” travel site which has really impressed me in my first day as a user. It replaces Dopplr (which I have used for approximately four months now) as my favorite startup travel destination on the web for two major reasons: its superior input methodology and the practical usefulness of the site’s main service: itinerary aggregation.

While these sites are clearly competitors, I think they might find that if a collaboration agreement could be reached, the sum would be greater than the parts.

Primary Functionality:

Dopplr LogoIf you were to ask me what Dopplr’s primary purpose was – its raison d’être – I would say creating community around travel, particularly for frequent travelers. It notifies me when I will be in the same place as one of my friends (still hasn’t happened to date, but I like the idea) so that we might meet up and grab dinner or a drink, or perhaps to share travel plans and tips. It also allows users to share their ideas and expertise about the places the visit frequently with other users online. In other words, it is a site that’s all about community. Unfortunately, there isn’t much of one yet. Until it has gained the faithful participation of more of my friends and acquaintances (which it has certainly been doing in the last few months), it just isn’t very useful to me.

TripIt, on the other hand, has a pretty compelling service from the get-go. It offers to aggregate the disparate elements of my travels into a single master itinerary. In effect, it does all those nice things your assistant would do for you when planning your travel, if you were lucky enough to have one. It allows me to look to a single location for all of my travel details: what flight I am on, what time it leaves, when it arrives, what seat I am in, what hotel I am staying at (at what address, with what phone number), and which rental car company I will be using to get there. It even provides a few handy “value-adds” such as weather forecasts for the locations I will be in each day, and quick access to city maps.

In January, TripIt added some social functionality and is attempting to build a community element which appears similar to Dopplr. Nevertheless, it’s community appears to be even thinner than Dopplr’s, and has a long way to catch up.

Input:

Whereas Dopplr offers a fairly easy and intuitive method of inputting travel locations and dates, TripIt introduces an input methodology that is truly groundbreaking (in my experience, at least). Instead of requiring any real effort on my part, all I have to do is forward them my confirmation emails (from United, Hertz, and Sheraton, for instance) and it parses the information to identify all the pertinent details. It loads this detail into my calendar instantly and automatically, even capturing things like my frequent flyer numbers.

Another blogger who recently compared Dopplr and TripIt suggested an even better idea: setup an email filter to automatically forward travel plans to TripIt, eliminating even that minimal effort required to put the site to work for me. With an email filter in place, TripIt would automatically aggregate all travel details, update my travel calendar, and stream it through iCal to calendar programs like Google Calendars. (As a side note, am I the only person who wishes you could use an iCal stream as an input into an existing Google Calendar entry, rather than requiring you to establish a separate calendar for external feeds?)

Once in my Google Calendar, my travel plans (and location) would be easily shared with friends and colleagues. Even better, once they join the TripIt community, we can even build collaborative itineraries (such as a business trip with several colleagues making arrangements for the group individually).

The Case for Collaboration:

In summary, TripIt has quickly won me over on its practicality and simplicity. Where it still falls far short of Dopplr, however, is on the community element. Dopplr’s Facebook application and blog widget (which I use here as well as at mitchellwfox.com) allow me to quickly and easily allow others to track my location. The potential value of discovering that a friend’s travels will overlap with my own is strong enough to convince me to continue updating my itineraries there in the meantime. If, however, TripIt’s itinerary aggregation and input could be joined with the powerful potential of community I see in Dopplr’s model, it would be a match made in frequent-flyer heaver.

Where from Here:

It will be interesting to see how the TripIt business model develops. In my initial usage of the site, I didn’t see any obvious indicators of what their eventual business model would be. Following in the steps of the likes of TripAdvisor by adding advertising and the ability to book trips would seem a logical course of action. One interesting suggestion made by another blogger was to enable travelers to re-book previous itineraries through a simple interface asking for the dates of the repeat booking, which could then be executed through a partner, such as Expedia or Kayak.  Given the convenience this would provide the user, you might be able to extract a small booking fee.

It is an exciting time in the development of online travel tools – I wonder what’s next.

Continue reading

Advertisements

Growing Excitement Around Product Recommendation Software

9 Apr

Richrelevance Logo

When you are shopping, a sales person who can quickly understand your needs, preferences, and budget and make a reasonable, logical recommendation is invaluable. While shopping online has typically required that customers already know what they were looking for, or that they conduct extensive research online in advance of a purchase, software is increasing playing the role of the sales person. While approaches to providing customer shopping recommendations have evolved with time, however, today’s software leaves considerable room for improvement. The recent funding of a Bay Area startup focused on customer recommendation demonstrates that venture capitalists have started to wake up to the potential this technology could hold.

Consumer recommendation tools online initially began by mirroring something that already existed in the print world: editor’s reviews and “product of the year” comparisons. Later came “Buyer’s Guides” which followed a simplistic logic to evaluate a few short responses to an online survey to provide a customer recommendation. Then came Amazon‘s product recommendations, based upon the analysis of other customer decisions (“others who purchased this item also purchased…”). This basic methodology has since been implemented in a number of different places around the web, with varying success. I would argue that NetFlix has been the most successful – the one site where I have significant confidence in the accuracy of the recommendations I receive, and act upon them with little or no knowledge of the film recommended. NetFlix, unlike today’s iTunes or Amazon stores, however, does this by not only considered what I have purchased (or viewed) before, but also how much I liked it.

If other online stores were able to earn my trust to a similar level without requiring the lengthy initial interview NetFlix used to gauge my movie taste, and were able to more deeply understand my shopping parameters, tastes, and the reason I arrived at their site, they would stand to gain a greater share of my wallet. If Amazon had been able to successfully recommend a book to me which I enjoyed (rather than assuming the South American literature textbooks I bought for college courses indicate a passion for Spanish authors), I would be far more likely to trust their recommendations a second time, and to begin to rely upon this functionality, visiting their store on a consistent and regular basis.

Baynote, a software firm located in Cupertino, raised $10.75 million in a second round of funding last year from Steamboat Ventures.  Its software attempts to understand customer intent by observing their actions on a website, and groups him or her into one of several customer archetypes to best deliver their anticipated needs.  The challenge, however, is that a customer’s visit may be so short as to fail to give enough evidence of intent for the software to accurately predict their intent.

Richrelevance, a San Francisco based technology startup which yesterday announced it had closed a Series B round of investment valued at $4.2 million dollars backed by Greylock Partners and Tugboat Ventures, is attempting to deliver this kind of next-generation product recommendation software. Built by David Selinger, a leader from Amazon’s recommendations team, richrelevance promises the ability to enhance a web store by personalizing the shopping experience and providing relevant, high quality product recommendations. Unfortunately, however, its technology doesn’t appear to make any massive improvements upon the flawed system in place at Amazon.

Perhaps this shouldn’t be surprising, however. It turns out the challenge of substantially improving recommendation algorithms and technology is a very considerable one. Even NetFlix, which posed a large cash reward to the tune of $1 million for any person or team which could improve the accuracy of its prediction software by 10%, has been unable to meet this seemingly modest goal after over a year and a half.

I will watch with curiosity as other companies tackle this challenge. I believe it is a field with significant growth potential, and one where I would be excited to see more innovation and expansion.  In the meantime, reasonably talented retail sales people need not worry about losing their jobs… just yet.