Research & data · October 2026 · Public Square Analytics

Borrow What Works: The Case for Proven-Practice Discovery in Local Government Communications

Every week, thousands of cities, towns and counties try out post ideas in public, and residents vote on them with likes, comments and shares. That makes local government social media one of the largest open libraries of tested ideas anywhere. We call the habit of searching that library, finding what has worked for governments like yours, and adapting it proven-practice discovery. This piece explains why it works, drawing on fifty years of research on how governments learn from each other and on 1 million posts from 5,834 local governments in 2026.

In this piece: What the research says · Recurring series that work · Eight posts worth borrowing · Format reuse in the wild · Common is not the same as proven · How good ideas spread · What we see in our own subscribers · How to practice it · How we measured · References

What the research says about governments learning from each other

Political scientists have studied how ideas move between governments for more than fifty years, under the name policy diffusion. Four findings matter most for a communicator.

Local governments learn from their peers, and learning is only one of the ways ideas travel. Charles Shipan and Craig Volden followed antismoking rules across the 675 largest U.S. cities from 1975 to 2000. They found four distinct channels at work: learning from cities that adopted earlier, economic competition with nearby cities, imitation of larger cities, and pressure from state governments. Imitation was the most short-lived of the four, and larger cities were better positioned to learn (Shipan & Volden 2008).

Governments copy what succeeded, not just what is nearby. In a study titled Emulating Success, Volden showed that states adopted children’s health insurance policies that had worked elsewhere, and that they looked to states similar to themselves rather than simply to their neighbors (Volden 2006). Experiments with U.S. municipal officials point the same way. Officials who were skeptical of a policy became more willing to learn from another city when the policy’s success was emphasized (Butler, Volden, Dynes & Shor 2017).

Copying can also spread the wrong things. In Why Bad Policies Spread (and Good Ones Don’t), Shipan and Volden argue that good ideas win out only when results are visible, there is time to learn, and decision-makers have the incentives and expertise to use the evidence. Without those conditions, imitation and competition can carry weak ideas just as easily (Shipan & Volden 2021). Sociologists describe a related pattern. When organizations are uncertain, they model themselves on peers that look successful or legitimate, whether or not the borrowed practice improves results (DiMaggio & Powell 1983).

Communication itself diffuses. It is not only policies that travel. In newspaper coverage of smoking restrictions, the way an issue was framed in one state was predicted by what other states had already done (Gilardi, Shipan & Wüest 2021). Government social media spread the same way. Agencies began with informal experiments by a few enthusiasts and only later settled into shared norms and policies (Mergel & Bretschneider 2013). Dutch municipalities took up websites, Twitter and YouTube at different speeds that depended on their size, structure and existing tools (Faber 2022).

One more line of research ties this back to day-to-day work. In the North Carolina Benchmarking Project, the cities that actually used comparative performance data to improve services stood out in three ways: the kinds of measures they relied on, their officials’ willingness to embrace comparison with other governments, and how far they built the measures into their management systems (Ammons & Rivenbark 2008). Comparison does not help by itself. It helps when someone looks at what worked and acts on it.

For social media this suggests a simple discipline. Look at what peers are doing, judge it by results rather than popularity, and adapt the ideas that hold up. That is proven-practice discovery.

Recurring series: the clearest place to borrow

The easiest ideas to borrow are repeat styles, recurring formats a government runs every week or month under the same name. They are easy to spot in another government’s feed, cheap to produce, and residents learn to look for them. Research on European local government Facebook pages has long found that engagement depends heavily on the type of content and media a page posts (Bonsón, Royo & Ratkai 2015). Recurring series let you test a content type again and again.

We identified twelve common recurring series in 2026 posts from 5,834 local governments. For every post, we compared its engagement (likes + comments + shares) with the same page’s typical post this year. That way a small town’s page and a large county’s page are each measured against themselves.

The series at the top share a pattern. They are about people and places residents already know: a bus driver, a crew, an old photo of a familiar corner, a local store. They also give residents something easy to do, like congratulating someone, sharing a memory or guessing a location. Milestone posts are the standout. They are also among the cheapest posts a communicator can make, because HR already knows the dates.

Eight posts worth borrowing

Each of these comes from a government that runs the format as a recurring series. Each drew far more engagement than that page usually gets. They load live from Facebook, so you see them as residents did.

City of Greenville, NC (Greenville Area Transit): Milestones post on Facebook

City of Ellensburg, WA: Milestones post on Facebook

City of Clovis, CA: Throwback Thursday post on Facebook

City of Hazel Park, MI: Throwback Thursday post on Facebook

City of Corpus Christi, TX: Photo of the week post on Facebook

County of Santa Barbara, CA: Guess where post on Facebook

Town of Phippsburg, ME: Business of the week post on Facebook

Township of Millburn, NJ: Employee spotlight post on Facebook

Format reuse in the wild: governments say so themselves

You don’t have to infer the borrowing. Governments often say it outright: “we’re jumping on the trend,” “challenge accepted,” “inspired by our neighbors.” Three patterns from 2026 show how quickly a format can move from one government to the next, and how the best ones adapt it rather than copy it.

“2026 is the new 2016.” In January a social media trend invited everyone to post photos from ten years earlier. We found 36 local governments in 18 states joining in, 29 of them within the same three weeks (January 12 to February 1). Several used the identical opening line, “We heard 2026 is the new 2016,” including Hopkins, MN, Roanoke, VA, DeBary, FL and Temecula, CA. The more useful versions bent the trend toward their own message. DeBary used it to thank employees who had served the city since 2016. Wheaton, MN used before-and-after photos to show what its downtown streetscape project had changed.

City of DeBary, FL: Borrowed trend, own message post on Facebook

City of Wheaton, MN: Borrowed trend, own message post on Facebook

The AI caricature wave. In the first two weeks of February 2026, at least 16 governments in 9 states posted AI-drawn caricatures of their staff, council or officers: four in Georgia alone and four in Texas. Many said plainly where the idea came from. Hewitt, TX: “Everyone is posting caricatures.” Texarkana Police: “Everyone else is doing it and we didn’t want to be left out.” Madison, AL: “We jumped on the AI trend… Meet your caricature Council!” Franklin, TN’s police department used the format to announce its next Coffee with a Cop.

Texarkana, TX Police Department: Joining a wave post on Facebook

Borrowing from a named neighbor. Some governments credit a specific peer. The Town of Danville, NH launched a coloring page “jumping on the bandwagon with our trend-setting neighbors at Hampstead, NH Police Department.” The Township of West Orange, NJ announced it would plant 250 trees for America’s 250th anniversary, drawing “inspiration from a community initiative recently adopted in Ewing Township” (post). That is policy learning and communication learning in the same announcement.

Town of Danville, NH: Credit to a neighbor post on Facebook

Challenge chains. Formats also spread by invitation. Edinburg, TX’s mayor called out the mayor of neighboring Pharr to join the Capable Kids Red Cape 5K. Ten days later the City of Alamo posted “Challenge accepted, City of Pharr” and asked who it should challenge next. In Colorado, Loveland and Windsor traded a mayors’ challenge the same way. Each challenge hands the next government a ready-made post and an audience that is already watching.

City of Alamo, TX: Challenge chain post on Facebook

These examples are the diffusion research in miniature: a visible idea, a peer network that carries it, and the best adopters adapting it to their own goals rather than copying it.

Common is not the same as proven

Look again at the bottom of the chart. Adoptable-pet posts are run by 212 governments, more than use photo-of-the-week or guess-where posts, yet the typical one lands slightly below the page’s own typical post. Weekly tips and week-in-review recaps do worse still: about two-thirds of them fall short of the page’s usual engagement.

None of this means those posts are useless. A weekly recap can be the right way to keep council actions on the record, and shelter posts find homes for animals. It means that copying a format because many governments use it is imitation. Copying it because it worked for governments like yours is learning. This is the distinction the policy-diffusion research draws (Shipan & Volden 2021; DiMaggio & Powell 1983). Proven-practice discovery means checking the results before you borrow.

How good ideas spread from one government to the next

To see recurring series spreading, we followed the 4,129 governments in our data that posted in every month from January to September 2026. That fixed group means a rising line reflects governments taking up a format, not our coverage growing.

• Employee and crew spotlights were used by 244 of these governments in January–March and 426 in July–September, an increase of 75% in six months.
• Spread follows peers. In a given month, a government that had not yet run an employee spotlight started one 2.5% of the time if another government in its state already had, versus 1.4% if none had. The pattern is similar for milestone posts (1.9% vs 1.0%) and business spotlights (0.37% vs 0.18%). For the series that underperform (tips, day-of-week hashtags, trivia), we see no such difference.
• Adoption clusters by state. About 30% of active governments in Kentucky and Arkansas ran an employee spotlight this year, against 16% nationally. Our City Hall Selfie Day analysis found the same pattern. In Georgia, where the municipal association runs a statewide challenge, 40.7% of active governments took part, against 5.5% nationally.

These are descriptive patterns, not proof of cause. Governments in the same state share associations, conferences and news cycles that could explain some of it. But the pattern is the one diffusion research would predict: formats that work spread through peer networks, and formats that don’t, mostly don’t.

What we see in our own subscribers so far

We also wanted to know whether governments that use GovFeeds to learn from peers see results in their own feeds. In September 2026 we compared 18 subscribing governments with 617 similar governments: the same state, the same type of government, and between half and double the population. We looked at the 12 weeks before and after each one started.

• How often they post did not change (+0.2% per week relative to peers).
• Their typical post drew somewhat more likes and comments relative to those peers after they started. But the rise was already under way before they subscribed, and with this few subscribers the difference cannot be distinguished from chance in our most careful test.

So we are not putting a number on it yet. Telling a real effect of 10–20% apart from week-to-week noise takes far more governments with long before-and-after windows than we have today. We will rerun the comparison as more subscribers reach twelve weeks on each side and publish what we find, whichever way it comes out.

How to practice proven-practice discovery

1. Choose your peers deliberately. Governments learn best from places similar to themselves, not just nearby ones (Volden 2006). Look at governments of your type and size, in your state and beyond.
2. Judge by results relative to each page’s own baseline. A big city’s raw likes say little about what would work for you. Ask whether a post beat that page’s usual engagement.
3. Borrow the format, not the post. Clovis’s history series works because residents love Clovis history. Your version should be about your people and your places.
4. Run it as a series. Give it a consistent name and a regular day so residents learn to look for it.
5. Check your own results after a handful of installments. Keep what lands with your residents and drop what doesn’t. That is the learning step that separates borrowing from imitation.

How GovFeeds supports this

GovFeeds is built for proven-practice discovery. It collects the public posts of more than 5,000 local governments and lets you search them by topic and filter to peers like you. It shows which posts engaged residents relative to each page’s own baseline, and the GovFeeds AI drafts your version grounded in real examples, with every source linked. The monthly report then shows how your own posts are landing.

How we measured

Data: public Facebook posts from 5,834 U.S. local governments (cities, towns, villages, townships, boroughs and counties) published January 1 to October 9, 2026, collected by GovFeeds; engagement (likes + comments + shares) available for 1,010,485 of them. Series were identified by their names and hashtags in the post text (for example “employee spotlight,” “#TBT,” “photo of the week”). Posts mentioning a death, memorial or tragedy were excluded from every series. “Typical post vs page’s own” is the median, across a series’ posts, of each post’s engagement divided by its page’s median 2026 engagement. The spread analysis uses the 4,129 governments with posts in every month from January to September 2026; a government’s first use of a series in a month from March onward counts as a start, and “another government in its state already had” means at least one other government in that group and state had used the series in an earlier month. The subscriber comparison is a stacked difference-in-differences over 12-week windows, with a placebo test; full details are available on request.

References

Links go to the publisher of record.

1. Ammons, D. N., & Rivenbark, W. C. (2008). Factors Influencing the Use of Performance Data to Improve Municipal Services: Evidence from the North Carolina Benchmarking Project. Public Administration Review, 68(2), 304–318. UNC School of Government
2. Bonsón, E., Royo, S., & Ratkai, M. (2015). Citizens’ engagement on local governments’ Facebook sites. An empirical analysis: The impact of different media and content types in Western Europe. Government Information Quarterly, 32(1), 52–62. doi:10.1016/j.giq.2014.11.001
3. Butler, D. M., Volden, C., Dynes, A. M., & Shor, B. (2017). Ideology, Learning, and Policy Diffusion: Experimental Evidence. American Journal of Political Science, 61(1), 37–49. doi:10.1111/ajps.12213
4. DiMaggio, P. J., & Powell, W. W. (1983). The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields. American Sociological Review, 48(2), 147–160. doi:10.2307/2095101
5. Faber, B. (2022). A Tale of Three Technologies: A Survival Analysis of Municipal Adoption of Websites, Twitter, and YouTube. Digital Government: Research and Practice, 3(3). doi:10.1145/3559108
6. Gilardi, F., Shipan, C. R., & Wüest, B. (2021). Policy Diffusion: The Issue-Definition Stage. American Journal of Political Science, 65(1), 21–35. doi:10.1111/ajps.12521
7. Mergel, I., & Bretschneider, S. I. (2013). A Three-Stage Adoption Process for Social Media Use in Government. Public Administration Review, 73(3), 390–400. doi:10.1111/puar.12021
8. Shipan, C. R., & Volden, C. (2008). The Mechanisms of Policy Diffusion. American Journal of Political Science, 52(4), 840–857. doi:10.1111/j.1540-5907.2008.00346.x
9. Shipan, C. R., & Volden, C. (2012). Policy Diffusion: Seven Lessons for Scholars and Practitioners. Public Administration Review, 72(6), 788–796. doi:10.1111/j.1540-6210.2012.02610.x
10. Shipan, C. R., & Volden, C. (2021). Why Bad Policies Spread (and Good Ones Don’t). Cambridge University Press (Elements in American Politics). Cambridge Core
11. Volden, C. (2006). States as Policy Laboratories: Emulating Success in the Children’s Health Insurance Program. American Journal of Political Science, 50(2), 294–312. doi:10.1111/j.1540-5907.2006.00185.x

This piece combines published research with GovFeeds’ own analysis of public posts. It is not itself a peer-reviewed study.

Frequently asked questions