What Factors Contribute to Attendance in College Football?

By Alex Koenig

A while back, John Ezekowitz and I tackled the question of attendance in the NBA and whether or not there are “must-see” road teams. Not surprisingly, our study showed that the opponent’s winning % was the most important factor – i.e. marginal fans came to see good teams. But that’s over an 82 game season (41 home games) in a league with only 30 teams. What about college football? With its 12 game season and 119 teams each home game is – generally – an event. Yet the average college football game between 2008 and 2011 only saw 80% attendance. Meanwhile, in 2010, 13-1 Nevada played in front of home crowds filled to 65% of capacity and 5-7 Texas sold out every game. What, then, drives college football attendance?

At the request of Winthrop Intelligence and with the help of Ed Gunther at The National Championship Issue I parsed through three years worth of attendance data to find out. Here are some of my findings, as well as some suggestions on how schools can improve ticket sales:


1. Sustained winning is important. The driving force behind big ticket sales is generally a compelling and successful team. Between 2008 and 2010 the average NCAA stadium was filled to 79.95% capacity – about 44,000 fans. However, teams that had winning seasons in that same time period drew crowds of around 53,000 (88.19%) – an 8% jump in attendance.

Teams that were coming off winning seasons – regardless of whether or not they maintained their winning ways – actually sold more tickets (89.03% of capacity) than their winning predecessors. This same difference is carried on, and exacerbated, to the upper echelons of success – teams with 90% or better winning percentages. It’s very hard to win 90% of your games, and fan bases reward that with home games for those teams averaging almost 63,000 fans (97.38%). But the reward is even greater the following season. Despite seeing their winning percentages drop, on average, to a more reasonable 73%, teams coming off huge years actually sold more tickets than they did in the middle of their success, coming extremely close – 99.98% – to selling out across the board.

Not surprisingly, losing has a similar effect on fan support. Teams that dropped below .500 played in front of crowds that weren’t even ¾ full, with an average stadium being filled to 70.5% capacity. The negative publicity that comes with losing continued to hurt teams the next year. Despite averaging win% hikes of 8%, teams coming off a losing season saw their attendance numbers drop to 69.7%. The lesson here is twofold: 1. winning helps as much as losing hurts; and 2. it usually takes a year for positive results on the field to be reflected with positive results in the stands.

Teams that have performed below .500 and maintained good attendance between 08-10

Colorado              Big-12                    36.1Wpct             0 weeks in top-25            92.7% cap.

Kansas State      Big-12                    48.5Wpct             0 weeks in top-25            94.5% cap.

Michigan              Big Ten                 40.2Wpct             0 top-5 conf. finishes      102.2% cap.

2. BCS Conference Schools draw substantially better. The typical (i.e., average) BCS conference school has 25,533 undergrads, one pro sports team in the area, wins 57% of its games and has a ratio of 55 regular citizens to every student in vicinity of campus. The average non-BCS conference school has 20,462 students, one pro sports team in the area, wins 46% of their games and has 59 regular citizens for every student. While those baselines are similar, the difference is reflected in attendance: the average BCS School attracts 61,000 fans to each home game, while the average non-BCS Schools attracts 24,000 fans to each home game. Although BCS schools having bigger stadiums (the same disparity exists as a percentage of capacity: 92% to 66%) and stronger on field performance (11% difference in winning% is significant but not that significant), the most likely reason for the disparity is the tradition of BCS schools and their opponents.

Non-BCS schools that avoided this problem between 08-10:

East Carolina      C-USA                   58.3Wpct             5-straight bowls                                98.0% cap.

Idaho                    WAC                      41.5Wpct             1 bowl in last 12 years    84.6% cap.

Southern Miss   C-USA                   56.4Wpct             9-straight bowls                                83.5% cap.

3. Undergraduate population does matter. During the 2008 to 2010 seasons, schools with between 30,000 and 40,000 undergrads averaged attendance of 54,000 (NCAA average: 44,000) which was 84% of their capacity (NCAA average: 80%) even with an average winning percentage hovering at just 50.6%. Schools with more than 40,000 undergrads – admittedly, a select few – had 69,000 fans at their games accounting for 91% of capacity. These bigger schools were also, generally, located in bigger metropolitan areas, and both had a citizen to student ratio of roughly 35 (NCAA average: 56.6). Smaller schools did considerably worse; schools with between 10,000 and 20,000 undergrads filled their stadiums to 73.2% capacity while those with less than 10,000 yielded 77% capacity returns, despite winning 57% of their games. These small schools (bellow 10,000) also were generally in large metropolitan areas and had a citizen: student ratio of 326.

Small Schools that have avoided this problem between 08-10:

Boise State         WAC                      94.9Wpct             19,993 students (28 ratio)             102.4% cap.

Navy                      Independent     67.4Wpct             4,400 students (115.8 ratio)         97.2% cap.

Oregon                 Pac 10                   81.9Wpct             16,475 students (20.5 ratio)         108.9% cap.

4. Large City population and the presence of pro teams hurt attendance. As much as a big school helps attendance, a big city hurts attendance. Medium sized college towns (100,000-500,000) filled their stadiums to 83.7% of capacity (50,000 per game). The same success was shown in larger college towns (500,000-1 million) with 83% attendance (44,000) and small college towns (less than 100,000) 80% (33,000). All this despite sporting average football teams (all hovering around .500). Colleges in cities (1 to 4 million people) attracted less than 40,000 fans to their games, only 75% attendance – below the NCAA average. The performance was worse in the biggest of cities (4 million+) whose colleges outperformed their competitors on the field – 58% winning pct. – but similarly only drew 40,000 fans (70% of capacity). One factor is the presence of other outlets for sports fans in these cities – i.e. Big Four sports teams. Colleges who did not have to compete with Big Four teams filled their stadiums to 83% of capacity, regardless of winning percentage. Teams that did? Just 74%. That’s a 9% difference, equivalent to 30,000 fewer tickets being sold per year by schools with pro teams in the area.

Schools with Pro teams in the area that have avoided this problem between 08-10:

South Florida     Big East                 61.5Wpct             Tampa, Florida (3 pro teams)      100.3% cap.

Cincinnati            Big East                 68.1Wpct             Cincinnati, Ohio (2)                          96.2% cap.

Utah                      Mountain West                84.6Wpct             Salt Lake City, Utah (1)                   100.8% cap.

5. Private schools have a harder time drawing. Even though private schools, on average, have slightly better football teams than public schools (a win% difference of +1%), public schools draw 5,000 more fans on average per game (80.5% vs. 76.5% capacity). There are three primary explanations for this:

  1. Private schools draw from a national pool of students. Therefore, there are fewer in-state allegiances passed on from generation to generation
  2. Private schools are smaller than public schools: Average of 25,259 undergrads at public schools compared to 10,791 at private schools. This means there is less value in undergraduate support and there is a smaller alumni base of fans.
  3. Private schools are generally located in bigger cities, closer to more pro sports teams. The average private school is in a metropolitan area in excess of three million citizens with just over two pro sports teams. For public schools, on average, there is less than one sports team and just over one million people.

Private Schools that have avoided this problem between 08-10:

Notre Dame               Independent     55.1Wpct             8,371 students (37.99 ratio)         100% cap.

Wake Forest              ACC                        42.6Wpct             4,476 students (108.3 ratio)         99.4% cap.

Brigham Young          Mountain West                71.7Wpct             32,955 students (14.39 ratio)       98.7% cap.

Athletic Directors and coaches can’t help the size of their school or what conference they’re in (2010 conference reshuffling not withstanding) and while they all are trying to win, each year half the teams will have winning records and half will have losing records. But one way they can influence attendance is in the scheduling of Non-Conference opponents.

As we’ve shown, BCS Conference teams draw significantly better at home than non-BCS schools. This is also true on the road. Between 2006 and 2010, BCS teams that took to the road for non-conference games filled stadiums to 90.83% of capacity, compared to just 83.9% for non-BCS schools. Notice that both of these numbers are significantly higher than the NCAA Average for that time period (79.95%). For whatever reason, fans show up for non-conference games in greater numbers than for in-conference games. Also of note, though it might hurt a team’s strength of schedule, hosting FCS schools only drops the average attendance to 78.71%, only slightly more than a 1% difference compared to the NCAA average.

Some notes on Non-Conference Scheduling:

1)      BCS Schools are reluctant to travel. Indeed, since 2006, only 379 non-conference games have been played in which a team from a BCS Conference was on the road. To put that in perspective, non-BCS schools played 650 road non-conference games and FCS (Division 1-AA) schools played 426. Of those 379 games, only 145 were at non-BCS schools. That means that out of the 1074 non-conference games involving BCS teams; only 13.5% did not have a BCS Conference host.  When these BCS teams played at non-BCS venues they saw crowds filled to 85.3% of capacity; an almost 20% increase over the mean for non-BCS schools – on average that’s 10,000 tickets per game. BCS @ BCS achieved 94.5% of capacity, an increase of 3% over average BCS-Conference ticket sales.

Not only are BCS Schools reluctant to travel, but they perform well on the road as well. Teams hosting BCS-Conference schools only won 48% of their games. That number drops to 34% when there’s a non-BCS host. From a cost-benefit analysis perspective (assuming a BCS-Conference team can be convinced to travel to your school) the value of the increased attendance has to be weighed against the high probability of defeat. That being said, a victory over a high profile can do wonders for a programs visibility.

2)      Non-BCS Schools have a rough time on the road. Home teams hosting non-BCS schools in non-conference games won 77% of their games. Despite dropping to a more reasonable 55% when a non-BCS school hosts a non-BCS school that is still a relatively high winning percentage. Coupled with 83.9% attendance (71.8% for non-BCS @ non-BCS) you’re not sacrificing much in the way of potential home defeat and not receiving as large of a reward in ticket sales if you choose to attempt to schedule non-BCS schools for your non-conference games.

3)      When it comes to FCS schools, try and maintain proximity. The conventional wisdom behind hosting a FCS for one of your non-Conference games is that it’s – hopefully – an easy win, and it can help serve as a tune-up for the rest of the season. To a large extent this is true. Since 2006, teams hosting FCS schools have won 92.5% of them, and those teams that hosted have had an average winning percentage of 54%, meaning the difference between an early season win over a FCS school and a loss in a more competitive non-conference game can be a bowl appearance. People also assume that scheduling FCS schools will result in low attendance (as we showed above) but this doesn’t necessarily have to be true. The average NCAA game fills the stadium to around 80% of capacity. Since 2006, more than half (54%) of games in which teams hosted FCS schools have achieved that number and more. Part of this is that these are usually the first games of the season (88% coming before week5), and with a fresh season comes enthusiasm from the fan base. But another important factor is careful selection of which FCS schools to invite.

Of the 64 games that sold out between FBS hosts and FCS opponents, the vast majority involved in-state, or bordering state, FCS opponents. This isn’t only true for powerhouses like Ohio State and Oregon hosting Youngstown State (102.65%) and Portland State (108.28%) respectively, but also non-BCS schools like Louisiana-Lafayette hosting in-state rivals McNeese State (133.41%) and Southern University (109.12%) – La-Lafayette’s average attendance over this time span was 61.8% – or SMU hosting Stephen F. Austin (108.59%) – SMU averaged 66.1% attendance. Getting the community interested in non-conference games is a great way to increase ticket sales.

4)      Cost-Benefit relationship for non-conference games (each case is different, but these averages provide a reasonable base line):


Difference from season averages are in parentheses

For BCS-Conference Schools:

Host BCS                              94.5% capacity (+3.01%)                48.0%Wpct (-8.69%)

Host Non-BCS                    89.6% capacity (-2.03%)                 87.8%Wpct (+31.1%)

Host FCS                              86.7% capacity (-4.9%)                   94.5%Wpct (+37.8%)

For Non-BCS Conference Schools:

Host BCS                              85.3% capacity (+19.4%)                34.4%Wpct (-11.4%)

Host Non-BCS                    71.8% capacity (+5.9%)                  55.0%Wpct (+9.2%)

Host FCS                              66.9% capacity (+1.0%)                  89.5%Wpct (+43.7%)

College football programs generally bank roll their colleges less popular athletic teams by making huge profits. In fact, the Huffington Post recently reported that some Football programs have yearly profits in the range of $60 million. Part of that is from merchandise sales, a lot of that is from TV deals, but the bread and butter of profits for college football teams comes in the form of ticket sales, and so – though the first priority should always be competing at a high level and providing an education for their student athletes – athletic programs should look to find ways to bring fans into the stadium.

About the author


View all posts


  • If your metric for high attendance is % capacity filled, I would respectfully submit that you’ve neglected a pretty crucial explanatory factor: playing in a small stadium. In fact stadium size may be the single most important factor, in that it explains the denominator completely, while the items on this list are all just contributing factors to the numerator.

    Most of the schools listed as exceptions to the poor attendance categories here play in stadiums that are small by FBS standards. Cincinnati, Wake Forest, Boise St, Navy and Southern Miss all have capacities in the 30k-35k range, well below the FBS average. The Idaho Vandals’ Kibbie Dome holds 16k, making it the smallest stadium in the division.

    Why not just look at attendance straight-up? Stadium size would seem to be driven by many similar factors to attendance, so dividing one figure by the other may confuse the result a bit.

  • You briefly mentioned it, but I think an important metric to look at would be alumni base. All things being equal, having an alumni base of 400,000 vs. 200,000 could be a huge advantage. This is especially true if it is a localized alumni base– like at a large state school.

  • Interesting graphs, but it appears that you used a linear trend on one and a non-linear on the other. Also, adding R-Squared calculations would go a long way towards establishing credibility for your assumptions.

  • Very nice post. I just stumbled upon your weblog and wanted to mention that I have truly enjoyed browsing your weblog posts. After all I’ll be subscribing on your rss feed and I’m hoping you write once more soon!

Leave a Reply

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