Is Antitrust Law Appropriate to Regulate Google and Facebook?

Recently Makan Delrahim, Assistant Attorney General for the Antitrust Division of the U.S. Department of Justice, publicly criticized Google and Amazon for their ruthlessness and size, actually and provocatively citing past government breakups of big private utilities like Standard Oil. If you’re surprised that someone in Trump’s DOJ would poke big business and invoke the progressive antitrust regulation of the past, some explanation is in order. 

The government has always regulated the market, so “free markets” have always been a kind of mythical creature. What governments typically do is regulate “competition,” but that doesn’t just mean keeping firms from becoming too big or stopping them from establishing monopolies through sheer size alone. It also means regulating certain behaviors, if those behaviors limit the choices or distort the autonomy of consumers. This is what’s behind the recent drive to use antitrust regulation to control big infotech. 

For example, digital advertising has both crowded out local and independent news, and intruded on consumer privacy. Both of those effects could be justification for antitrust regulation because that private data can be used in ways that disadvantage competitors who do not mine such data. Digital advertising is also a ripe target for regulation because this year, it’s expected to “exceed TV and print advertising for the first time ever,” and the Trump reelection team “is doubling down on its digital ad strategy,” according to CNN.

Just three companies–Google, Facebook, and Amazon–account for 70 percent of spending on digital ads, so that short list is another antitrust red flag. 

There is increasing public pressure to apply antitrust law to these big data giants. The public doesn’t like the Faustian bargain that has been made—search the internet and connect with others all you want, largely for free, but the companies get to surveil you “across the whole web” and use that data in any way they want. It’s like liberty is being traded for the right to communicate and gather knowledge. And in the meantime, competitors like Yelp have complained that the access power of Google has been used to crowd the smaller players out. Traditional data append and email vendors now represent just a sliver of the consumer data market.

“Outside of Google, Facebook, and a few others, the rest of the market, which includes thousands and thousands of independent news publishers (that depend on digital advertising as their primary source of revenue), will shrink by 11 percent” this year.

Sometimes size prevents competitors from developing. This can be true even if the size doesn’t translate into higher costs for the consumer. In fact, the largest tech companies today allow most of their services to be used for free, so we can’t really rely on price to test the assumptions of antitrust law. The data those companies gain in the process is really the problem, because the ability to take that data (which other companies have no access to precisely because those companies give their services away for free) and “identify untapped and under-served markets, spot potential competitors and prevent them from developing – the kind of edge that antitrust law is meant to thwart.” Antitrust regulation may also be warranted because those large companies create “natural monopolies.” Now, in the past, some natural monopolies were seen as acceptable as long as they were subject to additional antitrust scrutiny, like “price controls and oversight boards.” One could argue that Facebook, which is very close to being a natural monopoly, ought to be subject to a special government oversight board just for it. 

The European Union hasn’t wasted any time or parsed out any nuances here. Since 2010, the EU has investigated Google for antitrust violations three times and charged Google with violating EU competition law with Google Shopping, AdSense, and the Android system. The Google Shopping and Android charges stuck, and the company has handed over €8 billion to the European body. 

Despite this scrutiny (or maybe because of it), Google is acting impetuous and bold. Just a few days ago, the company, fully aware that it’s under antitrust scrutiny, announced that it was buying another company, the data analytics firm Looker, for $2.6 billion. It’s hard not to surmise that the company is testing the government to see who blinks.

Big Data, Business, and Politics

Some candidates in the 2020 Democratic presidential primary are marketing themselves like products, running solely on familiarity—presumably like the familiarity of your favorite neighborhood restaurant or a carmaker whose models you keep buying over the years. Others are using a radically different approach: They are building movements while they campaign, because they see themselves as activist leaders, not products. So these campaigns are building movements, “thoughtfully and deliberately designed to create an unprecedented grassroots movement driven by hundreds of thousands of volunteers.”

We hear and read these all the time: shallow comparisons between running businesses and running political campaigns—particularly where the use of data, social media, and other infotech are concerned.

It’s not entirely unfounded. After all, big political firms do make many of the same mistakes that businesses make where tech and data are concerned. For example, while I might not put it in the stark terms that he does, there’s something to Igor Lys’s comments in a recent post that many of the promises of big data in politics, just as in business, are false, based on the assumption that “big data allows reliable prediction.” I do think it’s futile to “predict” outcomes and instead that it’s better to use data in combination with other forms of information gathering. Lys agrees, writing that “the real use these people make of massive data collection and analysis concerns less the prediction and the manipulation of the future result, than the better analysis of the already existing ones.”

Political campaigns might also parallel business practices in efforts to capture email addresses of visitors that don’t end up donating money (or in business parlance, buy the product). Such sites might use pop-up windows to offer free newsletter subscriptions in exchange for an email address and use email append and verification services to build powerful databases of supporter contacts and preferences.

But while both political campaigns and businesses analyze data and collect contact info, I am also cautious about drawing too many parallels between, on the one hand, an endeavor whose primary goal is to make profits and one whose primary goal is to engage people into voting for, financially supporting, and working for political candidates or issues.

Here’s why political engagement, even through data use, is different from profit-seeking: Profits are extracted from workers’ labor and paid to owners or shareholders. These are very exclusive dividends. But political support grows as relationships among people. Political support, and political solidarity, are not finite and can’t be exclusively owned or claimed. A group of volunteers for a campaign may feel that political energy growing inside of them and when they share it with others, that energy grows rather than thins out. I don’t have less of it when I give it to you.  

This is why it’s important to use data in combination with direct political participation, such as social media engagement, canvassing, and campaign communications. For example, you can use your data to plan solid social media messaging strategies, to nuance the language on a candidate or issue website, or to pick the appropriate language for your campaign emails. Those messages invite different kinds of interaction, from campaign volunteers reaching out to vocal supporters of a campaign on social media, to a web form offering many different options for a supporter’s participation. Businesses sell products and services and the entire process is rather binary: will you buy the thing, if yes, then profit into the hands of owners and shareholders. There’s not much else a loyal customer can do beyond buy more products and refer others to do the same.

None of this is to say that pouring massive amounts of money into big data operations will have an effect even if it doesn’t help build an organic movement. Michael Bloomberg’s plan to use big data to help defeat Trump is a notable example of an effort that will probably have an effect even if it doesn’t empower people politically beyond voting. But an approach that includes participation and lots of interaction produces organizations —like Bernie Sanders’ campaign— that have strength beyond their numbers.

And, although I recently wrote that small donor acquisition efforts rely on creating a sense of urgency (which some may see as similar to creating consumer desire), these efforts are really ultimately about creating political communities. Sure, we can tailor social media ad campaigns based on appeals to different interests and demographics, but the end goal is to get invite these people’s participation through small donations as alternatives to courting large sums of corporate or millionaire money. And such courtship of small donors almost always includes inviting them into the interactive and participatory aspects of a campaign.

This isn’t some hypothetical or abstract philosophical assumption. Key voting blocs for the Democrats want big ideas and morally sound positions from candidates going into 2020. Knowing the difference in ethos between a political campaign and a business selling products or services is critical in appealing to the values of those voters who will create a new majority in the coming decades.

A Deeper Look at Segmentation

A Deeper Look at Segmentation

The main takeaway from a discussion of segmenting is to stop sending the same emails to everyone. But there’s another important takeaway for those doing digital political campaigns: people are not all the same, everybody brings unique and important perspectives into your campaign, and acknowledging that goes a long way towards getting supporters to help out in the best ways they can.

Segmentation originally emerged as a marketing strategy for email advertising of products and services. So it feels a little awkward talking about marketing strategies in relation to digital political campaigning, especially since many of the campaigns we work on are quite a distance away from the values of corporate advertisers. Two things get me through the incongruency. First, I don’t see an alternative in this system to raising money, doing so quickly and efficiently, and utilizing some level of mass marketing to do it. At the present time in the present world, it costs money to elect candidates who will defend reproductive rights, resist deregulation, and—if we demand it and support the right candidates—ultimately move us away from a pay-to-play political system.

Second, I’ve seen, and been a part of, thoughtful campaigns that use emails as part of larger conversations, of which raising money is an honest part, but not the whole. Strategies like segmentation sound cheesy, and in many contexts they can be cheesy. But strategies to find the kind of campaign supporters you want to engage are also essential, and if done right, can be the opposite of depersonalizing.

Scholars of persuasion going all the way back to Aristotle have emphasized the importance of understanding the different needs and situations of audiences. While political constituents have many common values and beliefs, they may have many different ways of getting there. Good communicators understand not only those common values and beliefs, but particular people’s unique starting points and priorities.

Even really market-oriented consultants say that good segmentation is about thinking beyond where people are right now. In politics, that means that people will reveal their policy hopes, and their aspirations about the political culture they want to live, if your email engagement can ask the right questions. Surveys are obviously an effective way to do this.

arrangement-blur-blurred-background-1181772.jpg

The process actually begins a couple of steps earlier. Email verification and appending for your voter lists is really important before you start sending pre-segmented emails that are designed to reach everyone on the voter list, because that’s the point where you’re most likely to have a lot of those emails land somewhere they don’t belong, whether that means bouncing or end up in the inboxes of those who don’t want them. So you’ll want to use an email verification, and email append with Accurate Append (including their lead validation) to get the best quality and coverage. Verification isn’t very expensive; it’s about a penny per email, and is worth it to avoid inbox penalties from email service providers.

From there, once you have an accurate list, you can email out a survey about issues in the district, with a simple landing page with little campaign branding so that people won’t feel like you’re baiting-and-switching them. The data from that survey will be like gold to your campaign team, allowing prioritization of subject matter for subsequent emails, social media engagement, and even traditional media engagement. Questions about voters’ income levels and ages can also inform both subject matter prioritization and fundraising, and you can also get that data from your vendor.  

Segmentation was one of the tactics in the Obama campaign team’s impressive use of email. Bloggers back in 2012 wrote enthusiastically about the campaign’s “detailed segmentation” that included sending different messages and asks to different levels of donors, or enthusiastic versus mild supporters. Committed supporters would receive a video, while undecided voters (yes, the campaign segmented them out too) received a chart showing job growth in Obama’s first term. Whatever might be said of Obama’s policies, he had a superior campaign team that made supporters feel like they were part of a larger vision. Every election, I work with candidates who have what I believe are critical and timely visions of policy. I want to know and respond to any information I can get about voters in those races.

While money is important, we know spending the most money doesn’t win elections (although the candidates who spend the most money usually win – it’s complicated, read the article). I think one thing that does win, when races are close, is candidates going the extra mile to understand and appropriately respond to their supporters, because doing so will inspire those supporters to put more energy and value into the campaign. Smart and thoughtful campaigns will append the gaps in their voter lists, gather specific information about the voters through surveys or other techniques, and then use segmentation to ask the appropriate voters for the appropriate things. None of that needs to be cynical or cheesy if you don’t want it to be.

Budgeting Your Digital Field Program

Here in California, voter ID programs are aided by a very complete file – phones, emails, addresses. In most voter ID programs, I also work with Accurate Append (client) to supplement voter files. If I don’t have emails or need more coverage, an email append can fill the gaps. Accurate Append’s services are essential for line-typing phones, where different contact methods are required for cells vs. landlines, and for validating emails.

Once your data is clean, using a simple calculator to estimate voter ID costs and frequency per contact method will help you determine where to deploy volunteers and how much you’ll need to spend to reach your goals.

Recently, I’ve been working on cost-analysis for voter IDs, and my firm is also beginning to use response data from our programs to create goals and budget for the meaningful interactions that precede a hard ID and help determine the health of a campaign. The sample calculator below includes social ads, bulk mail, and cost-efficient products from some of my favorite voter outreach solutions: CallHub for computer-based texting and dialers, VoterCircle for an influencer peer-to-peer program, and Ecanvasser for app and paper-based door-to-door canvassing. There are several notes on the original spreadsheet linked below the embed.

Voter ID program results are highly variable based upon scripts and targeting. Green numbers are variable and will change the results and costs totals when updated (you’ll need to copy the original spreadsheet). If you have suggestions or feedback, please email me, adriel@adrielhamptongroup.com. Hope you find this helpful!

Original Spreadsheet

Growing a Congressional Campaign Email List: Append vs. Acquire

One of the biggest challenges in running a Congressional race is keeping up the required fundraising pace. Email fundraising can help take some of the pressure off the candidate, but how to grow your list?

After starting with friends and family, it basically comes down to append vs. acquire. Here’s why I think a good email append strategy comes out slightly ahead here.

Append

Voter files often have some email coverage in place – the California file includes email addresses, and software like the VAN and PDI have email coverage but may charge fees to use those addresses (you’ll need a campaign-friendly bulk email system no matter what you do here – MailChimp will not allow the kind of voter prospecting I’m outlining).

For purposes of this exercise, let’s say you have about 300,000 voters on your Congressional voter file. An email append from a professional vendor like Accurate Append (client) will give you back deliverable email addresses for about one in three people on file. These emails come from consumer opt-ins throughout the web – however, they haven’t signed up for your list.

This list of 100,000 emailable people will cost you about $6,000, and may include data sweeteners like additional phone numbers for your voter file, and address standardization and correction using the Postal Service’s National Change of Address database.

What do you need to do to get these folks onboard?

First, I email a survey about issues in the district – one a landing page with little to no candidate/campaign branding, if possible. Typically this gives me an open rate of about 22% and perhaps a few hundred people take the survey. I included a prominent “leave this list” note with an unsubscribe link, also letting the prospect know I am reaching out because they are a voter in the district. Perhaps 2,000 of the original 100,000 unsubscribe.

Initially, I will be able to email about 20,000 of these folks, maybe more if I try additional survey tactics – however, I keep a very light touch to keep spam complaints low. I’ll begin to ease these 20,000 openers who didn’t unsubscribe onto my list. And for fundraising, I’ll bring in about $300 per email ($15 per thousand emails). I’ll estimate a that about half the list unsubscribes over the first couple months – but by that time, I should have brought in about $6,000 with 20 emails, paying for the cost of the list. Now I have an asset of about 10,00 voters who will still give about $150 per email, or $1,500 per month. Start early and this is a solid way to grow your list – especially if you’re reaching these local emailable voters with volunteer appeals and generally recruiting their support.

Acquire

An email acquisition strategy generally means partnering with a company that does this full time, or else investing a lot of ingenuity into a petition strategy. I co-founded Really American to help with large-scale signature opt-in acquisition efforts, and will say that getting district-specific emails is difficult and expensive.

But for this exercise, let’s say you can get 6,000 opt-in emails for your district for $6,000. ($1 each is a low estimate here.) These folks opt-in through a petition with your logo on it – but that doesn’t automatically make them happy subscribers! You’ll need to warm them up using Action Network Ladders or some other form of welcome series.

As you begin emailing these folks, you may raise twice as much per thousand vs. the appends, $30. That’s $180 per email, and if you send 10 a month, it’ll take just over three months to recoup your cash investment. You’ll also lose subscribers, ending up with perhaps 4,000 at the end of the first quarter of emails. That’ll be ongoing revenue of about $1,200 a month.

Append or acquire, there’s no easy money – but nurturing an email program can certainly pay for itself while achieving many of your outreach goals. Please read my article in Campaigns & Elections on content that nurtures a list to find more ideas.

Winning from the Ground Up

The new CTO of the DNC has already notched up a big win in Virgina, where the Democrats and aligned “WinVA” poured staff and new distributed field tools into the House of Delegates race, winning net +15 seats (three seats are still too close to call – in the balance are control of the Legislature and deeper influence on the state’s redistricting process).

I’m very pleased that VoterCircle played a role in this victory. Back in 2015, after being laid off by NationBuilder, I joined VoterCircle as an advisor – its growth has been extremely strong, but what’s most promising is VoterCircle’s potential to dramatically reduce the cost of elections. The software works by matching a campaign supporter’s personal network (email, generally, but LinkedIn and phone contacts also work) with eligible voters. At enterprise rates, it costs just $0.10 per contact – much cheaper than mail, and each contact is a confirmed email open. Whenever I prepare campaign plans these days, I lean in on VoterCircle – can we cut the cost of voter contact and then focus volunteer time on doors and phones with voters we cannot reach through a personal network?

Putting a digital touch on field contacts not only gives us a better look at who’s getting our campaign messages and how they are reacting, VoterCircle also learns from the campaigns networks, identifying the most influential voters with a campaign’s target universe. Campaigns can then focus special attention on supporters who can reach more than 10, 50, or more voters with just one email.

I also like to use cell phone and email data appends with VoterCircle to improve the quality matches (you can also use that data in survey-based email campaigns and in texting programs).

The most challenging barrier to adoption of a tool like VoterCircle is user trust – the DNC went a long way there by enlisting well-known surrogates:

With trusted, well-known surrogates, you’ve reduced the barrier to use – if you don’t have a Rosie or a VP candidate, try using the candidate themselves!

Wired magazine covered VA as an opportunity for the Democratic Party to take more risks in districts it may not win.

Tech donors bundlers also said they will be using data from VA to assist other Democrats use their money efficiently.

Read DNC CTO Raffi Krikorian’s thoughts here.

The Email and Contact Data Append API

Whether you’re running for city council with a vote goal of 2,200 people, or for President with a goal of 65 million, your campaign needs the best contact data to increase the number of connects to those voters.

Mailings can cost up to $1 each – every address had better be right.

For targeted calls, the challenge is connect rate. Say an unappended file has 30% coverage for phones. But it’s likely a third of those will be outdated! So you start at one-in-five voters contactable by phone. If you can increase coverage to 50% correct numbers, you’ve increased your contact rate 2.5x. That’s game-changing.

If you’re using commercial software and voters (or donors!) don’t have a “click-to-add” button for contact data, ask your vendor to implement the Accurate Append contact data API. You can then append landline and call phone numbers, emails, and update addresses if needed. Accurate Append has a basic implementation of the Open Support Data Interface, which means other adopters of that data standard can easily add Accurate Append data to their voter outreach tools and CRMs.

If you’re doing your own voter or donor database, access to the Accurate Append API is free for up to 500 calls for testing. With billions of data points – including hundreds of millions of emails and cell numbers – you don’t want to miss out.