It might seem like a matter of time before the NFL and London stop flirting and start going steady. Six NFL teams have flown across the Atlantic to play a football game this year — most recently, the Dallas Cowboys and Jacksonville Jaguars, who squared off on Sunday in the third and final London game of the season. (Just as if they were playing on home soil, the Jaguars lost badly.) The league would have to inconvenience only two additional teams1A 16-game regular season would require eight visiting teams to travel to London. to host a franchise in London full-time.Most commentary on the possibility of a London NFL team has been skeptical. Bill Barnwell, of Grantland, worried last year about travel and timing logistics and the potential disadvantages a London franchise would face in recruiting free agents.My view is more optimistic, at least when it comes to whether a London team could find a sufficient fan base. I’m not sure a franchise in London would be a smashing success. But even given conservative assumptions, London’s huge population and revenue base are probably enough to outweigh the relatively low level of NFL interest there. Perhaps more important, in contrast to some U.S.-based candidates for expansion or relocation, a London team would not cannibalize much of the fan bases of existing NFL franchises.Still, if London got first dibs on a team, the NFL would be overlooking a couple of more obvious candidates much closer to home.Last year, I looked at the National Hockey League’s allocation of franchises, estimating the size of each market’s NHL fan base using the population of its metropolitan area and the number of Google searches for the term “NHL.” (The analysis concluded that the NHL is overextended into smaller U.S. markets while underserving Canadian fans.) Here, I’ll perform the same analysis for the NFL, comparing cities that already have a team to potential new markets in North America and Europe.As with the hockey analysis, I’ll assume the popularity of the NFL in a given market is proportional to the number of Google searches for NFL-related topics,2The distinction between Google search topics and search strings is explained here. Topics are more comprehensive — for instance, Google searches for both “NFL” and “National Football League” will be grouped under the same topic. However, I default the search string for the term “NFL” in countries where topic-level estimates are not available, adjusting them upward to account for the less comprehensive search coverage. as according to Google Trends. Google searches might not be a perfect measure of popularity but they correlate reasonably well with other measures of franchise success3In the NHL, for instance, our Google-based estimates of each team’s fan base correlated strongly with its profitability. and allow us to compare domestic and international markets by the same standard. The only ad-hoc adjustment I’ve made is to lump Green Bay together with Milwaukee for purposes of calculating the Packers’ fan base.Otherwise, this is pretty simple: We’re just multiplying a metro area’s population4Market sizes are drawn from the estimates of metro-area populations put together by Demographia earlier this year. In the past, I’ve preferred to use estimates based on TV market sizes, which are slightly more inclusive of outlying areas that have a cultural affinity with a particular metroplex. But these aren’t widely available outside the U.S. and Canada. by the volume of Google searches it conducts on NFL-related topics. The estimated number of fans in each market is calibrated to the U.S. national average of 28 percent of Americans who say they are “very interested” in the NFL. Our estimates of the number of NFL fans in the 30 existing NFL markets5There are 32 NFL teams, but the New York and San Francisco-Oakland-San Jose metro areas have two teams each. — and about two dozen plausible expansion destinations — follow6Google Trends data is less detailed in some other countries than it is in the U.S. For cities in Canada, Germany, Spain and Mexico, I use data taken from the state or province level rather than the metro area. For London, I use data from the city level rather than the metropolitan area.:In contrast to the NHL (or college football), the level of interest in the NFL is fairly consistent from place to place in the United States. There’s also relatively little difference between those markets that have an NFL franchise and those that don’t.In some ways, these are signs of the league’s success: The NFL has conquered Sunday afternoons in just about every nook and cranny of the United States. And it’s principally a television sport. In the NFL, it’s not quite as important where the franchises are located — so long as you can transmit a TV signal from there.But partly because of the NFL’s pervasiveness, it has run out of highly attractive American markets other than Los Angeles. (Other than that, Mr. Goodell, how was the play?) Even Los Angeles provides some evidence of the league’s saturation: NFL interest there is only mildly lower than the national average despite the city not having hosted a team since 1994. Let’s say, however, that the NFL comes to its senses and places a team in Los Angeles soon. Where else is there to go in the U.S.?Las Vegas has high levels of NFL avidity and ranks as the next-largest untapped U.S. market by the number of NFL fans. But given the NFL’s longstanding paranoia about associations with gambling, putting a team there would be as much of an adventure for the league as going to a foreign market.After this are a series of markets — Orlando, Florida; Sacramento, California; Virginia Beach, Virginia; San Antonio; Austin, Texas; and Columbus, Ohio — where a team would play in the shadow of a more established franchise: The San Francisco 49ers in the case of Sacramento, for instance, or the Dallas Cowboys in the case of San Antonio. We’ll seek to measure the effect of this in more detail later on. It’s not that these markets are necessarily any less NFL-worthy than, say, Nashville or Jacksonville. But they’d reshuffle existing fans around more than they’d allow the league to expand its footprint.The foreign markets are more intriguing. Let’s start with London.I estimate from the Google data that only about 4 percent of Londoners are NFL fans now. However, the city’s metro area has about 10 million people. That means it has about 400,000 NFL fans. That isn’t great, but it’s comparable to a few existing NFL markets (Cincinnati, Indianapolis, Charlotte, Kansas City) and slightly larger than a few others (Buffalo, New Orleans, Jacksonville, Nashville). A London franchise might be the equivalent of a “small-market” team — but it would hardly be a huge outlier.There are a number of reasons to think this underestimates London’s potential. London is wealthy, with a GDP per capita of somewhere around £37,000 ($60,000 at current exchange rates). That means higher ticket prices and more billionaires to buy the team when it goes up for sale. London is also among the most-visited cities by tourists in the world with about 15 million international visitors a year.7Assuming that the average tourist visit lasts three to four days, that means there are about 150,000 international tourists in London at any given time. That’s small compared to London’s baseline of 10 million permanent residents, but it’s a nice little bonus. And it doesn’t account for travel there from within the United Kingdom, which is also significant.More important, our estimate that 4 percent of Londoners are NFL fans is based on the volume of Google searches since 2004. Those searches have increased recently, and there’s reason to expect a further increase in fan interest if a team is located in the city permanently. As measured by Google searches, interest in the NHL increased by about 80 percent in the province of Manitoba, Canada, after the league relocated a franchise to Winnipeg in 2011. The NBA experienced a similar increase in Oklahoma City when it moved a team there.8It also helps that the Oklahoma City Thunder have been much better than their predecessors, the Seattle Sonics. Because the NFL is already so saturated in the United States, I wouldn’t expect an 80 percent increase in NFL interest if you placed a team in Orlando or Austin. But London, and other foreign markets, have a much lower baseline and more room to grow.A London-based team could also have some appeal across the rest of England and the United Kingdom. One precedent comes from the Toronto Blue Jays and Toronto Raptors, the only Canadian teams in Major League Baseball and the NBA, respectively. Each one generates about 20 percent to 25 percent as much search traffic in other Canadian provinces as it does in its native Ontario. That doesn’t sound great, but it’s higher than most U.S.-based franchises, many of which generate only about 5 percent as much search traffic outside their home states. With no other franchise to compete against geographically, a London team could be regional in the way the Denver Broncos, Dallas Cowboys and Boston Red Sox are, covering a larger footprint than you’d infer from its metro area alone.You might think these are pie-in-the-sky assumptions; I think they’re pretty reasonable. The only issue is that there are two other international destinations that rank better still.They’re not among the more exotic choices. Paris, Dusseldorf9Essen-Dusseldorf, which also includes several other mid-size cities, is the most populous metropolitan area in Germany. and Madrid almost certainly would not have the fan bases to support an NFL team at the present time. A second U.K.-based team, in a place such as Manchester, would not do much better. Nor in all likelihood would San Juan, Puerto Rico, which is a baseball town.But the Toronto metro area is highly populous and NFL interest is already reasonably high there. I estimate T Dot has about 1 million NFL fans — more than the majority of U.S. markets to host an NFL team. As with the Raptors and Blue Jays in their sports, there could also some residual gains in NFL interest across the rest of Canada.Mexico City ranks even higher. Although only about 7.5 percent of people there are NFL fans, 7 percent of 20 million residents is still 1.5 million NFL fans.Could those Mexico City fans afford tickets and licensed replica jerseys and the products sponsors might want them to buy? Mexico gets pigeonholed as a developing country and that’s true for much of the nation, but Mexico City itself has developed into a thriving, bustling city with many of the creature comforts available in the other great metropolises of North America. Mexico City’s metro-area GDP is about $30,000 per capita and GDP per capita is nearing $50,000 in the city proper, comparable to that in U.S. cities. Levels of NFL interest in Mexico City, while not extraordinarily high, are higher than in London: An NFL game there in 2005 drew more than 100,000 spectators.The international markets also offer the advantage of being unconquered territory rather than existing in the shadow of any current NFL team. To measure this, I ran another series of Google Trends searches on topics related to individual NFL teams (e.g. searches for topics related to the Seattle Seahawks) to see how they compared to interest in the NFL as a whole.In existing NFL markets, Google search traffic for the local team is generally about 65 percent to 70 percent as high as that for the league as a whole. See here for the Detroit Lions, for example. Of the Detroit area’s roughly 840,000 NFL fans, Google search volume would suggest we’d allocate about 480,000 of them to the Lions. Another 200,000 or so would go to the next-most popular NFL teams there, the Green Bay Packers and the Chicago Bears. That leaves relatively few “free agent” fans.In the foreign markets, however, including in Canada, fans are largely not committed to any one NFL franchise. In the table below, I’ve estimated the number of fans for the three most popular teams in each market and calculated how many fans remain after allocating fans to those teams.10The calculation is a bit rough for some of the less promising markets. Interest in the NFL is low enough in Dusseldorf that we don’t have a great idea of who the most popular teams are there. Also, in England, the most popular team according to Google topics is nominally the Cleveland Browns. But this appears to be a false positive, with Google having picked up on other contexts in which the word “Browns” is used. Searches for the text string “Cleveland Browns” as opposed to the topic “Cleveland Browns” are quite low in England. In Mexico City, for instance, the Cowboys, Pittsburgh Steelers and Broncos are probably the most popular teams. But searches for those three teams combined represent only 20 percent to 25 percent of searches for NFL-related topics as a whole. Contrast that with Columbus, where searches for the Cleveland Browns, Steelers and Cincinnati Bengals represent about 90 percent of searches for the NFL as a whole. That’s not to say a Columbus-based team wouldn’t pick up some fans of its own, but they might come largely at the expense of the Browns, Bengals and Steelers rather than acquainting new fans with the league.Toronto, like Mexico City, has only about 20 percent of NFL fans allocated to one of the three most popular NFL teams there. The Buffalo Bills have sometimes protestested that Toronto is part of their market, but NFL fans in Toronto take only a modest interest in the Bills according to search data and other metrics like merchandise sales.I estimate there are about 50,000 Bills fans in greater Toronto. That isn’t nothing when there are only about 300,000 NFL fans in metro Buffalo itself. But that’s Buffalo’s problem, not Toronto’s. If the NFL wants to have a franchise in Buffalo, it should have one in Buffalo. It should also have one in Toronto. The league would come out ahead if it had to slightly subsidize the Bills with the extra revenues it gained from a Toronto team.How about Montreal or Vancouver instead? If you could combine the virtues of the two — Montreal’s larger population with Vancouver’s greater NFL interest — you’d have an NFL-worthy city. As it stands, however, both are decidedly inferior to Toronto. Montreal comes out slightly better than Vancouver in our reckoning because, while each has about the same number of NFL fans, a fair number of those in Vancouver are committed to the Seattle Seahawks.Among U.S. cities, Los Angeles remains No. 1 with a bullet after allocating fans to existing teams. Las Vegas’s numbers also hold up well. So, to a lesser extent, do Orlando’s, a surprising result given that there are three other NFL teams in Florida. But Orlando, like other cities in the state, has a lot of expats from the north who root for teams like the New England Patriots and New York Giants and who might or might not be intrigued by an expansion team. The state of Florida has produced its fair share of disappointments in cultivating loyalty toward new franchises. Most of the other American candidates could wind up like Jacksonville — at best just barely big enough to support a team on its own and with that team having barely any footprint beyond the city’s borders.A final question is about the NFL’s endgame. If the NFL merely needs a couple of credible candidates for relocation — whether as leverage against existing teams or as genuine alternatives — Los Angeles and London should more than suffice. But if the league is thinking about expansion, it might have to do it in a big way. Thirty-two teams is a convenient number, readily divisible into two conferences and eight divisions of four teams each. A 33-, 34- or 35-team league would be awkward, however. The next equilibrium would be 36 teams instead, which could be divided into six divisions of six teams each.In that case, the NFL ought to return to cultivating the Mexico City market and treat Toronto as more than a token alternative for the Bills. An expansion to those cities along with London and Los Angeles would be the boldest thing the league has done in years — and possibly the smartest.CORRECTION (2:15 p.m.): An earlier version of a chart in this article misstated the number of unallocated NFL fans for Mexico City. That number is 1.15 million, not 1.49 million.