Yes, 2016 will be noted as the year in which the Internet came into its own in U.S. political campaigns. While it long has been the technology for the back-end operations of political campaigns, this year online media have broken through as one of the primary sources for information about the election.

Throughout the primaries, social media often eclipsed paid media and coverage by traditional news organizations. Donald Trump’s tweets often became the story, drawing all media coverage to his message and disarming his opponents without relying on significant fundraising. As the general election begins, some analysts project that Internet political advertising will exceed $1 billion.

That projection seems optimistic to me. One billion dollars would represent a shift from about 1 percent or 2 percent of all political advertising spending in 2012 to more than one-fourth in 2016. Campaign techniques rarely change so quickly. That said, online advertising will continue to grow. But how does it fit with political campaigns?

First, elections in the United States are deeply connected to geography. They are conducted at the levels of states, municipalities and districts. Campaigns need to target areas — states and congressional districts — as much as certain groups of people. Winning the presidency hinges on winning a subset of key battleground or swing states. Winning control of the U.S. House depends on which party wins in 40 or so swing districts.

The inescapable geography of elections means that communications technologies that can be tailored to geography are most effective to use. Television media markets are efficient for some states, such as Florida, but very inefficient for others, such as New Hampshire, because reaching New Hampshire voters requires advertising in the Boston media market while reaching Florida voters can be accomplished entirely by advertising in Florida media markets. And for most congressional districts, TV is very inefficient.

Geography is, and will likely continue to be, a significant limitation on the use of online advertising in elections. Online advertising, at least as we know it today, is not very effective at targeting voters in key geographies, such as a state, let alone a congressional district. It seems more likely to get a toehold in campaigns for the presidency, especially during the primaries, where candidates seek to create a national buzz around their campaigns.

Second, we don’t yet know how effective online advertising is in persuading people, either generally or in politics. The study of the effectiveness of televised political advertising is thin, and more often than not shows that political ads have very small effects on viewers. Some studies of banner ads suggest that they are less effective than traditional advertising. There are problems of invisibility and clutter, and the medium is sufficiently new that the research on the effectiveness of the simplest political banner ads has yet to be conducted. It is unknown whether what we know about advertising from the television sphere translates to the online medium.

To the extent that there is a consensus among researchers about the effects of television advertising, it is that advertising reinforces viewers’ prior beliefs more than it mobilizes them to vote or persuades them to view politics differently. The rise of online advertising, then, will likely have a similar effect: reinforcing beliefs rather than changing minds.

In the early 1960s, V.O. Key famously described elections as an “echo chamber.” Voters know what they want but can only respond to the choices they are offered. In Key’s time, for good or ill, there was a common, national media. That chamber, first with the rise of cable and now online communications, is increasingly divided into thousands of distinct chambers and channels. What we are offered is increasingly targeted to what we want to hear and see.

Online advertising will continue the increasing segmentation of the American electorate. The distinct advantage of online advertising is the power to target (and avoid) very specific demographic and political segments of the public, based on audience characteristics of Web pages or the click behavior of individuals. Online advertising offers a far greater ability of candidates, parties and groups to reach only those people they want to reach, and to leave the others out of the political discourse.

Stephen Ansolabehere is a professor in the department of government at Harvard University.