I think the Sanders campaign has been too cautious with spending because they had literally no idea they'd get this much. 2,000,000 unique contributions from ~850,000 unique donators is absolutely unprecedented; it's ahead of Obama, it's ahead of everyone. And, you know, when faced with something like that, you'd think "maybe it was a fluke, we'll put some aside just to be safe" - but the rate of contributions has gone up, not down. Heck, DFA's endorsement is worth just over $1mil! And I think the Sanders campaign is realizing that - Sanders purchased his first set of ads in the last few weeks, and sources on the inside are saying the big guns are coming out soon. Until now, Clinton has had a massive spending advantage - she's spent $44mil to Sanders $13mil. Kind of unsurprising she was doing so much better, Sanders brought a knife to a gun-fight. Now the shoe is on the other foot (or perhaps, the gun is in the other hand?).
A few points on Iowa:
Sanders has actually been carpet-bombing Iowa and New Hampshire with TV ads since the beginning of November. He's actually
outspent Clinton in ads in the past month (dropping $5.5 million between November 12th and December 15th compared to Clinton's $4 million). The talk by his strategists back in early November was "watch out, this is going to make a difference!", but the polls have only showed Clinton's lead increasing since then.
Truthfully ads aren't going to make that much of a difference at this point anyway. Sanders has only 5% less name recognition than Clinton at this point in Iowa according to Selzer, and other polls like Quinnipiac report a similar story. You can argue that Sanders is at a name recognition disadvantage nationwide that increased press and ad campaigns could overcome, but Iowans are pretty clued in.
Overall Clinton's only outspent Sanders by $4.3 million in ads this year. All that money she's been burning through hasn't been invested in ads but in ground game. That doesn''t show up in the polls - it didn't for Obama in 2008 - but it absolutely makes a difference on election day. Clinton's got the very best - I'd argue at least as strong as Obama's 2008 Iowa operation. And it's too late for Sanders to compete at that level. Not only does it take months to fine-tune, but he hasn't got access to the very best staff, who are overwhelmingly in Clinton's campaign, who know how to game the complex caucus rules and find every last voter. She also has the vast bulk of local party activists on her side.
Sanders focus in Iowa these past couple of weeks:
Rallies as per. Clinton's focus:
Precint Leadership Training. The extensive focus on field will make a difference.
I don't think Selzer can accurately predict the turnout expected on caucus night (unless there's a source?) but she can pick up the percentage of first time caucus goers expected. She predicted that 60% of Democratic caucus goers would be caucusing for the first time in 2008, and despite the Clinton camp and Edwards camp calling that number absurd she was proved right (the number was 57%). So far she's not picked up numbers anywhere close to that figure. If Sanders really was likely to be expecting a flood of new voters to overwhelm Clinton on caucus night Selzer would be picking it up in the voter registration lists by now.
This isn't going to be about votes cast, it's going to be about the delegate math. Even Selzer's poll can only poll the former, which likely underestimates Clinton's lead since she'll have the advantage when it comes to the latter. Selzer's noted that Sanders support is more geographically clustered than Clinton's - most noteably in college towns and cities - which means he's going to be harder for him to pick up as many delegates statewide.
This is where the caucus date helps Clinton compared to 2008. Back then the caucuses were in January during the winter break, which meant Iowan college students were back at home all over the state. When they went to caucus the their votes were distributed fairly evenly in all 99 counties and 1682 precints, which gave Obama a big delegate advantage. This time in February they're going to be caucusing in a handful of college towns like Iowa City and Ames. And getting 10000 students to caucus for Sanders in a precinct won't get him any more delegates than 1000.
One other way this isn't 2008: it's essentially a two person race and that plays to Clinton's advantage. Clinton turned out around ~70,000 caucus-goers in 2008 - more than Edwards and not far off Obama - but she got pummeled in second round voting and ended up third. When Obama/Edwards/Richardson/Biden failed to meet the 15% threshold in a precinct, very few of their second preferences then went to Clinton. She was much much more divisive in Iowa - her unfavorable ratings among Democrats in Iowa was near 25% (thanks to her Iraq War vote), it's now closer to 10%. There's not much evidence O'Malley's 4% is going to break heavily one way or other. So this year it isn't about reaching out and building alliances, it's about raw turnout, and Clinton has the edge - both demographically and organizationally.
Truthfully I don't know whether the polls are going to tighten as the caucus gets closer, or whether Clinton will continue to build on her lead (as voters run to the safe option like they did in 2004). But she's led every Biden-less poll but one (internet poll) in Iowa all year (c.f. eight years ago when Clinton trailed Edwards for the first half of 2007, took a small lead in the summer, and then fell behind Obama after the JJ Dinner) and has a big organizational advantage that will make a difference of an added few percent on caucus night. External events might break things Sanders way, but I don't see how more ads will make up the difference. I think Clinton is going to take this by 15%+