Dave, While I agree with your conclusions, I think it is important to look carefully at your data... --On Tuesday, January 05, 2016 06:08 -0800 Dave Crocker <dhc@xxxxxxxxxxxx> wrote: > On 1/5/2016 4:44 AM, Christian Hopps wrote: >> I've often wondered if when polls were taken if any weight >> was given to active or long term contributors. > > Thanks for raising this. The surveys should, indeed, be > tailored to produce more useful information, with more > emphasis on pragmatics. > > The IETF meeting has a core of regular participants. They > dominate the surveys we currently do. They are well-funded > and well-traveled. ... and probably well-represented by the people on the Meetings Committee. > To be serious about efforts an inclusiveness, venues should be > easy, quick and cheap to get to and cheap to stay in. > > The surveys should primarily target folk who are /not/ > guaranteed to attend but who are nonetheless desirable > attendees. > > This requires better sampling -- don't just query current > attendees -- and better questions -- don't just ask about > general preferences for a particular city. Yes. And don't just ask about how well the last meeting went. > Our current sampling method produces a tourism focus. > > Jari's comment: > >> On 1/4/2016 12:27 PM, Jari Arkko wrote: >>> Out of the last fifteen meetings: >>> >>> Yokohama, Prague, Dallas, Honolulu, Toronto, London, >>> Vancouver, Berlin, Orlando, Atlanta, Vancouver, Paris, >>> Taipei, Quebec, Praque, Beijing >>> >>> I count only two (Honolulu and Orlando) that were clearly >>> touristy destinations. > > demonstrates some of the problem we have in considering cities > carefully. The reality is that most of those venues are > highly popular tourist destination, especially in summer. > > Four of them are listed as among the top 20 in one survey: > > http://www.forbes.com/sites/deborahljacobs/2014/07/31/the-20- > most-popular-cities-in-the-world-to-visit-in-2014/ But that story appears to be ultimately based on a Master Card analysis [1]that intermixes business and vacation travel (and other categories, if they exist). So I don't think one can get from that list to "popular tourist destinations". > Three of them, in another: > > http://www.tripadvisor.com/TravelersChoice-Destinations This appears to be a more tourism-oriented list. When one excludes locations (or even countries) with no significant Internet engineering presence from that list, those three are actually an uncomfortably high percentage. One could look at other surveys and other criteria. By some of them, the _only_ city on the recent meetings list that is not aggressively promoted as a tourist destination is Dallas. > When we travel to such places during the height of their > tourism season, we encounter bigger crowds and higher prices, > both in transit and in staying. Indeed. And, I would assume, hotels that are reluctant to allocate large blocks of rooms to us at discounted rates if they think they can make more money on the per-room basis from tourists. That leads to another conclusion: many locations have clear "low" and "high" seasons. They can be great meeting locations in the low seasons because hotels and other facilities are hungry and more anxious to accommodate (on rates, room blocks, and more generally) than when they are convinced that most of their rooms will be filled no matter what they do. Minneapolis or Toronto in the dead of winter or Phoenix in July might be good examples of that -- whether those are tourist destinations in the high season or not, they almost certainly are not in the low season. >... > Instead, we should query potential /additional/ or > /infrequent/ attendees who are showing up on discussion lists > already and who are not well-funded. > > The form of the surveys also should be different. Simply > asking for basic preferences about specific cities elicits a > response about the appeal of the city, not about the > pragmatics of going there. Unfortunately, if it is a city that few IETF attendees have visited before(especially to attend large work-oriented meetings) information about those pragmatics may be largely missing and those with information may be hard to distinguish from guesses by people who think the city would be interesting. > So the surveys should begin by priming the context by asking > general, policy-related questions about venue factors, such as > travel price and travel time and venue costs (including food) > and venue convenience (isolated resort versus resource-rich > urban environments). > > After that it should ask about specific locations but should > include information about each place's costs and convenience. > > And it should not just ask about preference. It should ask > about attendance likelihood. That is, cast the question so as > to elicit a mild form of commitment. People answer such > questions differently than open-ended preferences questions. That seems right to me. best, john