How to Spend A Billion Dollars

In all the discourse about how much the Harris campaign raised, how they spent it, and if that worked, there’s been surprisingly little coverage of the actual filed details of their financials. I happen to love FEC filings, and they’re pretty accessible online. Dealing with donation records can be time consuming and require more sophisticated data handling to avoid crashing your computer, since there are millions of records. You also run into things like donations under $200 cumulatively not being itemized, many donations flowing through ActBlue, etc. 

Lucky for us, spending (disbursement) data is far simpler. Here’s the summary of spending by the Harris for Presidential/ Kamala Harris For The People entities, helpfully provided to us by the FEC. 

This is all also available in their pretty decent data explorer, for your scrolling enjoyment. If you’re trying to look at extremely recent data, like the day a filing comes in, this isn’t the best interface because it takes some time for the FEC to process filings. For older data, it’s totally fine, which is what we have here. 

We can just straight up download the file of disbursements as a csv. It includes about 92k records, which will make an excel sheet lightly sad but not sad enough to bother with anything else. Each disbursement is a row, with information on what vendor it went to, the report year (by default, this includes spending from the primary, which is fine for this use case), amount, and a categorization of disbursement description. How clean this file is depends on the rigor of whoever is doing the filing from the campaign. Big national campaign files tend to be nicer, congressionals tend to be messier. This one is quite nice, thankfully. 

We need to do a little clean-up first- a lot of individual small disbursements are things like contribution refunds for people who went over the allowed dollar amount or requested them. I’m going to cut these out because they don’t represent an interesting piece of spending. 

Group by vendor, sort by spend descending, and boom, you’ve got a breakdown of where that billion dollars actually went.

Media Buys Are Expensive

The top 4 places money went are all media buys and production. Together, these 4 media firms represent 707 million dollars in spend, 61% of the total disbursements reported by the campaign. Note that this does not include any PAC spending, FutureForward is entirely outside the scope of this report. 

Because of the useful transaction labeling, we can look at media buys more completely. If you look at media buy, production, consulting, and paid media, you end up with around $809 million dollars, which is roughly 70% of total campaign spending. The vast vast majority of this is to a few big firms, trailing off into one or two thousand dollars at a time to local tv stations, papers, and so on. 

This all flows through media buying firms, so we can’t easily figure out how much money went directly into ad purchases and how much went to vendor overhead. In my experience, ad buyers (especially TV ad buyers) tend to take a percentage of the spend on ads as a fee, which is likely what’s happening here. I don’t do enough work with TV ads to have a sense of the exact percent, but something like 7-10% seems likely from basic googling. 

There’s a lot of unanswered questions buried in that media spend. Was it effective and efficient? Where exactly did these ads run? Is this too much money to be spending on ads? These are all difficult to answer without digging into records of what ads ran where, which requires a subscription to something like TVeyes which I don’t have.

One thing we can say is that *generally* it is most efficient for campaigns to directly buy and run TV ads. This is because of the lowest unit rate rule, which means that campaigns in the last 60 days of an election must get the lowest class of rate for a given spot. Since media gets so saturated in the last days of an election, this rate can be hugely cheaper than what an outside org would pay (see my comparison of rates in GA for an example). I’m not hugely bothered by this massive media spend flowing through the campaign for this reason, it was likely the cheapest option for the ads that ended up running. 

Top Spending Categories 

So, what are the top types of spending? We can categorize by disbursement_description and get a pretty okay view. Below are the top 20 categories of spending, by total amount spent.

Media is the biggest category by far, as we discussed above. 

Another chunk of spending (although FAR smaller) is payroll. 54 million is a lot, but it’s only 5% of the total. You can in theory look at payments to individual staffers here, but I find that ends up obscuring the issue for the very highest-paid staff. They tend to be getting funds through multiple methods, including consulting fees or payments to associated firms, and I don’t feel like the direct payroll records are really representative. Plus, it feels a bit creepy. 

Text message outreach

There’s several different vendors in here, including Action Network, Scale to Win, Action Squared Inc (which is just Action Network under a different name), and Switchboard, which isn’t unusual. They might be split out by type of text, or method of sending. The rules around bulk texting vary by the type of list you’re contacting, cold outreach vs opt ins, etc. 

The pricing on texts also varies, but we can do some deeply back of the envelope math by assuming it costs around 1 cent to send an SMS from a shortcode number (I got this by looking at ActionNetwork’s fees plus carrier costs). Assuming nearly all the cost here went to texts themselves (yes, there’s overhead probably), that’s something like 4 billion texts sent. Now, that sounds pretty high, so I might be mis-estimating the costs here, but it’s certainly a fuck ton of text messages. Modern campaigns rely heavily on text outreach for fundraising, as your messages in November probably showed, and some of this traffic is likely post-election fundraising efforts to fill their budget gaps. 

I think it’s interesting that this is showing up as a major line item for a presidential campaign. Even a few years ago, this wouldn’t have been the case, whereas media spending is a constant. 

Service fees

Why is this category so big? Turns out it’s almost entirely ActBlue service fees for using their platform to raise money. They charge 3.95% per donation in processing cost, and when you’ve raised a billion dollars, that adds up. This was a real windfall year for ActBlue. 
Egencia is a travel management company, so I’m guessing the service fees are a similar type of thing, some percent of money passing through them. 

Democracy engine is similarly a fundraising tool, although they appear to have used it FAR less than ActBlue. 

Polling

Finally, a juicy category. The bulk of this money looks to have been run through the DNC. Lucky for us, the DNC also does FEC filings! 

I’m not sure why some of this was run through the DNC and some was done directly to the pollsters. The rules around sharing information (things you’ve paid for can count as in-kind donations) get messy fast, and arrangements like this aren’t uncommon, but I don’t have insight into why they did this specifically. 

This is everything marked as “polling” on the DNC’s disbursements filing.

You may notice that 17 million is much more than the 7 million passed over by the Harris campaign. There’s not a good way to tell what pieces of this were paid for by the Harris campaign. Potentially you could do something clever and compare payment dates? But with lots of polling running all the time, this isn’t going to be easy. 

The most obvious thing is that a handful of vendors were the “big guns” here. Grow Progress is an ad testing shop- they let you run your ads in a survey and see which perform better- so I wouldn’t count that as traditional polling. All these vendors look extremely normal, and the biggest thing I have to say is “wow, that’s a ton of polling”. This makes sense for a presidential campaign, and I do personally love to have my field be well funded, but I also worry about if we’re reaching the point of diminishing returns. Although the boom times for fundraising are clearly over, so this is unlikely to be a problem in the future. Our tight budget future will be a whole new world. 

Event Production/AV/ etc

This is an incredibly messy category with a lot of vendors and a lot of different types of spend. Using the complex analytical method of “selecting anything with ‘event’ in the description” and “also things marked as audio visual services”, I see around 81 million dollars of total event spending to 618 different vendors. That includes even minor event expenses, like charges at the Home Depot and $199 to “Wristband.com” for, presumably, wristbands. 

There’s 19 different entities that got over a million dollars in event spending, from 9.8 million to “Wizard Studios North” to an even 1 million to Harpo Productions, which is Oprah's production company. Does this count as “giving Oprah a million dollars to endorse” as people claimed on twitter? Probably not, since it looks to be event production services and not an entertainment fee. Other fun bits include $375,000 to Three Wishes Productions, which is Christina Aguilera, and $187,000 to Kitty Purry Inc, which can only be Katy Perry. 

With a tiny bit of digging, the biggest vendors here seem to be mostly sound equipment and other basic venue equipment rental companies. This was probably mostly for rallies and similar events across the country. I don’t have a great baseline for how much events are “supposed” to cost, but renting the equipment to put a thousand people in a space and deliver a speech safely and audibly seems quite expensive for good reasons.

Canvassing

This is distinct from canvassing done by direct employees of the campaign and their recruited volunteers. Things falling in this category are likely what you’d call “paid canvassing”, which is contractors employed by vendors to knock doors. 

I am a known critic of paid canvassing, and I don’t generally think it’s effective or efficient. In the overall context of the campaign, I’m not wildly thrilled to see this line item, but this is just not that much spending. 8 million dollars in the abstract is a lot, but on a billion dollar campaign, it’s not huge, and I’m happy to see a diversity of tactics being tried. 

Where did the money go?

Ads! It mostly went to ads. So many ads.

This is a boring answer, but I don’t think most normal people understand just how much advertising was run by this campaign. We don’t get a breakdown by ad type in this data- some of this was likely digital, I can’t really tell how much- but TV ads in particular are very expensive when you’re blanketing the airwaves of every swing state. 

Some twitter discourse has called advertising in general a waste of money and called for a return to more field-based modes of campaigning. This is, frankly, kind of silly. However craven running infinite ads might feel, we know empirically that they work to convince voters. And we know from similar research that field programs are typically more expensive per vote and able to reach different (usually smaller) populations of voters. It might not feel as authentic, but modern campaigns happen in advertising and media, not at community meetings. There’s another discussion to be had about the decline in local democratic participation, the disappearance of local Democratic chapters, etc, but throwing money at things that aesthetically match what you want politics to be doesn’t bring those institutions back. We should try everything, but do what works. And so, yes, putting 60% of your budget into ads is probably a reasonable choice. 

The other strain of discourse has been that there’s too much money in politics, and specifically too much money in presidential campaigns. I am much more sympathetic to that, not least because I mostly work on downballot campaigns with comparatively teeny tiny budgets who could really use a few thousand extra dollars. I can’t begrudge the Harris campaign the massive initial fundraising bump they got when she entered the race. There’s clearly an appetite among donors to fund these campaigns, and to fund them to the extreme. However, campaigns can also get caught in the fundraising momentum and spend ever more money on trying to raise additional funds, hire more staff that then need paying, and so on. This clearly happened to the Harris campaign a bit, since they ended with a deficit. I don’t know that the solution is for a campaign to set a cutoff point, but there might be a level at which you should pivot away from fundraising and just plan for what you’ve already raised. 

I don’t think the next few years will sustain this level of money flowing into campaigns. Donors are clearly tired, Democrats are out of power, and we have already somewhat exhausted the threat of Trump as a fundraising mechanism. I would expect that the next few cycles of Democratic campaigns are going to have to re-learn how to operate on a limited budget, and I would further expect the Harris campaigns funding numbers to be a higher water mark for some time. It’s going to be a different world out there, and we should start bracing ourselves now. 

A side note on FEC filings

There is a ton of useful and interesting info buried in these filings. If you’re a journalist and you want to do more work with these, I am happy to chat. I also occasionally do more complete versions of this analysis as consulting projects. If you do decide to dive in, please take some time and read the FEC’s own rules on what gets disclosed and how before freaking out about any seemingly-illegal donations (I say this from experience).