From one in three to nine in ten

The message

In 2011 about a third of American adults owned a smartphone. By 2024 it was 91 percent. Adoption curves like this one rise fast and then flatten as they approach everyone. The design should show the flattening, because "almost everyone, and growth has stopped" is the current state.

Real dates on the x-axis

Pew didn't survey on a fixed schedule. Some years have several surveys, some have none. We placed each point at its actual survey date instead of spacing points evenly. Even spacing would make the gap between January 2018 and February 2019 look the same as the gap between February 2021 and September 2023, and the slope would lie.

We also picked one survey per period instead of plotting all of them. The fact sheet has more than thirty readings from 2011 to 2024, and early ones bounce around by a few points from survey to survey. Plotting every one would add noise without changing the story. The data table below lists the points used.

The y-axis runs to 100

Percent-of-population data has a natural ceiling, and showing the full 0 to 100 range makes the remaining gap visible. A tight axis from 30 to 95 would exaggerate the recent flattening and hide how close the line is to the top.

Labeled where it matters

Three points carry labels: the start (35%), late 2016 (77%), and the latest (91%). The big "91%" in the corner repeats the endpoint at poster size, because that is the number people will quote.

Survey numbers have margins

Every survey estimate comes with a margin of sampling error. Pew publishes its methods with each survey. For a chart like this, a difference of a point or two between neighboring surveys shouldn't be read as a real change, which is one more reason to plot a sparse set of points.