Growing Followers: Collaboration, Shares, Paid Reach and Why Fake Growth Costs You

The beginner course teaches how to read reach, engagement and follows (reading the numbers). This lesson is about the mechanism behind the follow count and what the evidence says moves it. Most follower advice is anecdote dressed as method. The parts that survive scrutiny are fewer and plainer than the industry suggests, and the one shortcut that always works, buying followers, is the one that costs you the most.

How a stranger becomes a follower

A person follows a page after seeing something from it. For someone who does not already follow you, that exposure arrives by one of two routes. Either a person they are connected to shared it (a reshare, a story share, a send in a direct message), or the platform recommended it to them without any connection (Reels, Explore, the For You feed, Facebook's suggested content, YouTube's home and Shorts feeds). There is no third route except paid promotion, which is recommendation you buy.

Both routes run on the same signals. Instagram's ranking explainer says the most important predictions for Reels are "how likely you are to reshare a reel, watch a reel all the way through, like it, and go to the audio page," and that for Explore the popularity signals, "how many and how quickly other people are liking, commenting, sharing, and saving a post," matter much more than in Feed (Instagram, May 2023). TikTok says a strong signal is whether a user finishes a longer video, and that follower count is not a direct factor in recommendation (TikTok, 2020). Facebook says recommendations favor videos with reactions, comments and shares beyond plays (lesson 05).

The 2024 Instagram change made the mechanism explicit for small accounts. Instagram said it would recommend every eligible Reel to a small group of users it thinks will be interested; if they engage, the Reel goes to a slightly larger group, and the cycle repeats. Only original content is eligible, reposts are labeled to point at the original, and content aggregators are barred from recommendation. Adam Mosseri said the goal was to "help small creators reach a broad audience," and acknowledged that large accounts previously had an edge from their initial reach (NBC News, 2024). For a page with 2,000 followers that means a reel's first test audience is strangers chosen for interest, and the reel either earns the next round or does not, on completion and reshares.

Likes are the weakest signal in every one of those statements. They are listed, but behind reshares, completion, saves and comments. A page that measures itself in likes is measuring the thing the systems weigh least.

Shares as the engine

The research on social spread is in lesson 02; two results do the work here. Bakshy and colleagues' Facebook field experiment with 253 million users showed that exposure to a friend's share made people significantly more likely to share, and that the many weak ties carried most of the novel information (arXiv:1201.4145). Bond and colleagues' 61-million-person experiment in Nature delivered to 61 million Facebook users during the 2010 congressional elections, found that the messages' effect through social transmission on real-world voting "was greater than the direct effect of the messages themselves," and that "nearly all the transmission occurred between 'close friends'" likely to have a face-to-face relationship (Nature 489, 2012). The two findings fit: close friends' actions move people to act, and weak ties carry the news to people who had not heard.

For a page, the practical consequence is that the thing worth optimizing is the reshare, and reshares follow from content people are willing to attach their name to. The Rival IQ nonprofit finding in lesson 05, that Facebook photos were shared three times as often as other types, points at documents and real images. People share proof and useful things. They like opinions. The growth is in the first.

Growth tactic	Mechanism	Evidence	Strength of evidence	Rule to respect
Make reshareable content (documents, useful explainers, real people)	Reshares reach non-followers through weak ties; reshare is a top ranking prediction	Bakshy 2012 (253M users); Instagram ranking explainer; Rival IQ nonprofit share data	Strong for the mechanism; benchmark for the format	Original content only for recommendation eligibility
Collab posts with partner accounts	One post appears to two audiences, adding weak ties	Instagram's own recommended method for sharing others' content	Platform statement; no independent effect size found	Real partners; X bans substantially similar posting across accounts you control
Commenting on larger pages	Your name appears before an audience that does not follow you	No controlled evidence found this session	Practice, not science	No coordinated comment rings; no bait; add something
Live video and events	Notifications to followers; higher comment volume	Facebook 2016: "more than 10 times more" comments on Live than regular video, from "initial data"	Platform claim, old and preliminary	Idaho: an unpaid live is not reportable; a paid promotion of it may be
Series formats	Returning viewers; a reason to follow rather than consume once	YouTube retention reports split new from returning viewers; no causal study found	Measurable on your own page; no external effect size	None
User-generated content	Supporters' posts carry your message through their ties	Bond 2012 on friends' displayed actions; no verified UGC-specific effect size this session	Mechanism strong; tactic-level evidence vendor-sourced and omitted	FTC endorsement disclosure if anything of value changes hands (lesson 09)
Paid boosting	Bought recommendation to a chosen audience	WordStream, Superads, M+R benchmarks with disclosed methods	Observational benchmarks; your results will vary	Authorization, disclaimer, platform eligibility, Idaho reporting
Buying followers or engagement	None; the accounts never engage	Instagram 2018 removals; Pew 2018 bot share; every platform's policy	Strong that it is detected and penalized	Prohibited everywhere

Collaboration and cross-promotion

Instagram's Collab feature lets accounts "co-author content with other accounts," so one post lives on both profiles with one set of engagement (Instagram Help). Instagram's creator guidance names it as the right way to share a partner's work: "If you see a reel you think your followers would love, you can share it to your Story by using the share button," and when "working with another creator, use our Collabs feature to create a single post that is shared to both of your followers," while re-uploading content already on Instagram is discouraged and can cost an account its recommendation eligibility (Instagram for Creators). A coalition of three groups posting one explainer as a Collab reaches three audiences with one piece of original content, which is exactly the weak-tie gain lesson 04 describes, done in the way the platform rewards.

The failure modes are the ones lesson 04 catalogs. Partners who all repost the same video separately get the duplicate-content penalty. Accounts one person controls that like and repost each other violate X's rule against using your accounts to amplify one another and Bluesky's rule against manipulating engagement signals. Partners who agree to comment on each other's posts on a schedule are, on X, "trading or coordinating to exchange follows or Post engagements," which is prohibited (X policy). The honest versions are easy to tell apart from the dishonest ones: real partners, real posts, each account's own voice, and a Collab or a share with attribution rather than a copy.

Commenting on larger pages, the staple of growth advice, has no controlled evidence we could find this session. The mechanism is plausible: a substantive comment under a county newspaper's post puts your page name in front of readers who do not follow you, and a comment that adds a fact (the page of the budget, the date of the hearing) gets replies and visibility. Treat it as practice. Do it with one account, say something true and useful, never coordinate a group to do it, and never post the kind of "comment YES" line that Meta demotes.

Lives, events and series

Facebook's 2016 launch post for Live said that "from initial data, we've seen that people comment more than 10 times more on Facebook Live videos than on regular videos" (Meta Newsroom, Apr. 6, 2016). It is a decade old, preliminary, and from the company that sells the feature. What has not changed is that a live stream of a public meeting, with someone naming the agenda items as they come up, is content nobody else has, and that followers get notified. For a page whose theory of change runs through meetings, a scheduled live series from the meeting room is one of the few formats that is both original and recurring.

Series earn follows because they give people a reason to come back. The evidence for that at the page level is in your own analytics: YouTube's retention report lets you split new from returning viewers (YouTube Help), and a series whose returning share grows week over week is working. We found no external study quantifying the series effect for small civic pages, so the claim here is only that it is measurable and that a page should measure it.

Paid reach: what a dollar buys

Boosting a post buys recommendation to an audience you specify. The benchmark data with disclosed methods, all observational and all from advertisers rather than advocacy pages:

• WordStream analyzed 2,946 Facebook campaigns from February 1, 2023 to April 30, 2024, reporting medians: cost per click $0.77 for traffic campaigns and $1.88 for lead campaigns; click-through rate 1.57 percent and 2.53 percent; cost per lead $21.98 (WordStream).
• Superads, from more than $3 billion in Facebook spend across its customers from September 2025 to August 2026, says CPMs "range from $5 to $18 depending on the region and objective" in most industries, using medians to blunt outliers (Superads). The page does not give a single figure, so treat the range as the estimate.
• M+R's 2026 Benchmarks, medians across 180 participating nonprofits, report a cost per lead of $2.81 overall and $3.64 on Meta, with other lead vendors (Care2, Civic Shout) at $1.59 and display at $86; 71 percent of lead-generation spend went to social; fundraising return on ad spend was $2.48 on search, $1.11 on display, $0.56 on digital video and $0.04 on TikTok, where the cost per donation was $590 (M+R Advertising).

The reason to spend is specific: a document post that your own followers shared heavily is a candidate for a boost to an audience that looks like them, because the organic signal says it works. Boosting a post that did nothing organically buys impressions on something people already declined to share.

The rules that come with paid promotion

Paid promotion on social issues, elections or politics is the most regulated thing an advocacy page does, by the platforms and by Idaho. As of October 2026:

Platform	Can an advocacy page pay to promote on social issues, measures or candidates?	What it must do first	Disclaimer	Public record
Facebook and Instagram	Yes	Complete authorization (identity and location); provide organization details for the disclaimer; EIN, FEC ID or .gov/.mil credential for a Confirmed Organization name	Verified "Paid for by" required; ads without one are disapproved	Ad Library, seven years
Google and YouTube	Federal and state candidates, parties and qualified state measures: yes, after US Election Ads verification. Idaho local candidates and measures: no, "not allowed"	Verification; synthetic-content checkbox where applicable	Google generates "Paid for by" for most formats; advertiser must include it in some	Ads Transparency Center
X	Yes. Political content ads (referencing candidates, measures, legislation) allowed; political campaigning ads (advocating or soliciting) require certification, US buyers only, limited formats and targeting	Certification for campaigning ads	Advertiser must comply with applicable disclosure laws; no on-ad format specified	Political Ads Library disclosure report for campaigning ads
TikTok	No. Its ad policy has barred political ads from any advertiser since 2019; its guidelines prohibit "paid political promotion, political advertising, or fundraising" by politicians and parties and bar paying creators to support or oppose a candidate; politician and party accounts have ads turned off	Not applicable	Not applicable	None
Bluesky	No ad system	Not applicable	Undisclosed advertising prohibited; label any paid placement	None

Sources: Meta's 2019 rules on authorization and disclaimers and its current advertising standards (Meta Newsroom; Meta Advertising Standards); Google's political content policy, whose state restrictions list for Idaho reads "Ads related to ballot measures and candidates for local elections" as not allowed, alongside Maryland, Nevada, New Jersey and Washington with similar or broader blocks (Google Ads policy); X's political content policy (X Help); TikTok's 2022 political accounts post (TikTok Newsroom); Bluesky's guidelines (Bluesky). Meta also restricts new political ads in the final days before a US election; campaign finance lesson 06 has the current detail, and the practical rule is to launch early.

Google's Idaho block deserves a sentence of its own because it surprises people every cycle. A page working on a city council race, a school levy, a county bond or a library district measure cannot buy YouTube or Google search ads about it, at any budget, with any verification. State legislative races and statewide qualified measures are allowed after verification. Meta, X and direct mail remain open, which is why Meta carries most local political spending.

Then Idaho. The Secretary of State's Guidance 2026-5 says unpaid social media posts generally need no report and that "paid broadcast messages via radio, TV, text, or social media" are paid political activity (Guidance 2026-5). From there the thresholds in lesson 04 apply. A boost that expressly advocates for or against a measure or candidate, made without coordination, is an independent expenditure; over $100 for one race or measure it needs a 67-6611 statement, and $1,000 or more after the sixteenth day before the election needs one within 48 hours. A boost that refers to a candidate within 30 days of a primary or 60 of a general is an electioneering communication over $100 under 67-6628, and its statement names everyone who gave the spender $50 or more. Paid express advocacy must identify its sponsor under 67-6614A. A nonprofit whose own paid political activity passes $1,000 in the year files the nonbusiness entity statement covered in campaign finance lesson 08.

The operational answer to all of that is one habit: no boost without a line in the log. Page, date, amount, who paid, the text of the post, whether it names a candidate, whether it uses express advocacy words, and the election date. The treasurer reads the log weekly in season. Most reporting failures are not evasion; they are a volunteer who boosted a post for $60 on a Tuesday and another who did it for $50 on Thursday, and nobody added them up.

Follower quality: why fake growth hurts

Every platform prohibits bought followers and engagement, and every one of them says so in rule text. Instagram announced in November 2018 that it would remove "inauthentic likes, follows and comments" from accounts using third-party apps, notify the accounts, and warn that continued use "may see their Instagram experience impacted" (Instagram). TikTok's integrity guidelines bar trading in followers or likes and say that when it finds inflated metrics it removes the fake followers or likes (TikTok). X prohibits "selling or purchasing followers or engagements" and engagement exchanges (X). Bluesky prohibits artificially manipulating "engagement metrics, follower counts, or other measures of community interest" (Bluesky).

The scale of automation on at least one platform is documented. Pew's 2018 analysis of about 1.2 million English-language tweets linking to 2,315 popular websites, collected over six weeks in 2017 and classified with Botometer, estimated that suspected bots posted 66 percent of those links; political news sites were the lowest news category at 57 percent, and the bot share of links to sites with liberal and conservative audiences (41 and 44 percent) did not differ significantly (Pew). Pew called them "suspected" because Botometer is about 86 percent accurate by one estimate. The point for a page is that raw share and follower counts on X are not clean measures of human attention, yours or anyone's.

Why purchased followers lower reach follows from the ranking statements rather than from a published experiment, so take it as inference. Instagram lists among its Feed signals "information about the person who posted," meaning how often people have interacted with the account in recent weeks, and ranks by predicted engagement from each viewer (Instagram). Benchmarks compute engagement rate as engagement over followers. A thousand accounts that never like, comment, share or watch add to the denominator and to the pool of followers who are shown your post and ignore it. Every signal the systems weigh gets worse. The platforms then remove the fakes anyway, and a page that bought 5,000 and lost 4,000 in a sweep has a growth chart that explains itself to any reporter who looks.

The 90-day plan

Growth targets should come from published growth rates, scaled to your page, and then adjusted for how much of the work above you will actually do. The published rates:

Benchmark	Source and method	Annual growth	Implied 90-day growth (annual x 90/365)	Implied weekly growth
Instagram, 1,000 to 5,000 followers, 2025	Socialinsider, 35M posts, 447,613 pages; followers gained over followers at start	22.0 percent	5.4 percent	0.42 percent
Instagram, 5,000 to 10,000 followers, 2025	Same	20.3 percent	5.0 percent	0.39 percent
Instagram, 10,000 to 50,000 followers, 2025	Same	17.2 percent	4.2 percent	0.33 percent
Instagram, 1,000 to 5,000 followers, 2024	Same, prior year	38.0 percent	9.4 percent	0.73 percent
Nonprofit Facebook pages, 2024 to 2025	M+R, 180 participants, median; average page 53,000 followers	3 percent	0.7 percent	0.06 percent
Nonprofit TikTok accounts, 2024 to 2025	M+R; average 3,000 followers; the report gives 34 percent in one place and 37 in another	34 to 37 percent	8.4 to 9.1 percent	0.65 to 0.71 percent
Nonprofit X accounts, 2024 to 2025	M+R	minus 3 percent	minus 0.7 percent	Decline

Sources: Socialinsider Instagram benchmarks; M+R Benchmarks 2026, Social Media. Two readings before you plan. First, small accounts grow faster in percentage terms because the base is small; 22 percent of 3,000 is 660 people, which is thirteen a week. Second, Instagram growth for small accounts nearly halved from 2024 to 2025 (38 to 22 percent), so a plan built on 2024 rates would miss by half. Benchmarks age fast; check the current year's before you set targets.

Questions and answers

Our followers grew 400 in a week after a post went viral, then engagement fell. Did we do something wrong?

Probably not. A viral post reaches people who liked that one thing, and some follow and most never engage again, which lowers engagement per follower by arithmetic. Instagram's ranking depends partly on how recently each viewer has interacted with you, so those followers see less of you over time. Watch whether the next few posts' shares hold. If the 400 came with no shares at all and the accounts look empty, run the audit; bot waves follow viral posts too.

Is it worth boosting a post to get followers, or should every boost drive email signups?

The benchmarks only measure the second. M+R reports a median cost per lead of $3.64 on Meta; nobody with a disclosed method reports cost per follower for boosted organic posts, so you would be buying an unmeasured outcome. Boost for the lead, measure the follows as a side effect, and compute your own cost per follower from the log. If it turns out cheap, you will know; if it does not, you have the emails.

Can we run the boost from a volunteer's personal account to avoid the authorization hassle?

Meta's authorization applies to whoever runs ads on social issues, elections or politics, and the disclaimer will show that person's legal name if no organization is provided (Meta). In Idaho, the spender is whoever paid, and if the organization reimburses the volunteer, the organization paid. Do it properly from the organization's account, with the organization's name. The shortcut puts a volunteer's name on a public ad record for seven years and muddles the Idaho filing.

Our local school levy campaign cannot buy YouTube ads. Is that right?

Yes, as of October 2026. Google's state restrictions list says ads related to ballot measures and candidates for local elections are not allowed in Idaho (Google Ads policy). TikTok allows no political promotion. Meta allows it with authorization and a disclaimer; X allows it with certification for campaigning ads. Campaign finance lesson 06 covers the offline alternatives.

A vendor offers "real, active followers" for $50 per thousand. Why not?

Because every platform's rules prohibit it, Instagram has removed such follows since 2018, TikTok says it removes fake followers when it finds inflated metrics, and the accounts never engage, which lowers every signal the ranking systems weigh. And because the growth chart is public. A reporter comparing your follower jump with your flat comment counts has a story in an afternoon. The $50 buys a liability.

How many Collabs or partner shares is too many?

On X, the limit is in the rules: substantially similar content across accounts you control, or coordinated engagement exchanges, are violations regardless of count. On Instagram, Collabs with real partners are the recommended method and we found no stated cap, but every Collab is a post your followers see, so the ceiling is your audience's patience. One or two a month with different partners, each adding a different audience, fits the weak-tie logic. Weekly Collabs with the same partner are cross-posting with a label.

Does going live still get special distribution?

Followers can be notified, which is the main mechanism, and Facebook's 2016 claim of ten times the comments was preliminary and is a decade old. We found no current platform statement of a live-video reach advantage. The case for live from public meetings is that nobody else has the content and it recurs, not that the algorithm favors it.

We are a 501(c)(3). Can we boost at all in an election season?

You can boost on measures as lobbying within your limits, and on issues, with authorization and a disclaimer. You cannot boost anything that supports or opposes a candidate, and a boosted post that refers to a candidate inside the 30 or 60 day window is an electioneering communication under 67-6628 that would also be a federal tax problem. Campaign finance lesson 08 has the tax rules and the nonbusiness entity filing that starts at $1,000 of your own Idaho political spending.

Hard questions

Isn't follower growth a vanity goal? Why devote a lesson to it?

Because followers are the base of the funnel in lesson 08, and because the theory of change in lesson 02 needs specific people to see specific things before a vote. The lesson's answer to the vanity objection is in its measures: shares per post, leads per post and cost per lead, not the follower count alone. A page that grows followers who never share or sign up has grown a number. The 90-day plan's week-13 review is built to catch exactly that.

Doesn't paying for reach mean the platform, not the public, decides who hears you?

Yes, and so does not paying; the recommendation systems decide either way, as lesson 01 explains. Paid reach lets you choose the audience and makes you disclose who paid, in the Ad Library for seven years and in Idaho's filings. Organic reach lets the system choose the audience and discloses nothing. There are good reasons to spend little; the campaign finance thresholds and the owned-audience argument in lesson 07 are two of them. Purity about paid reach is not among them, since the unpaid route is also mediated.

The authorization and disclosure rules favor organizations with an EIN and staff. Isn't this a barrier to ordinary people?

Partly. Meta's Confirmed Organization label requires an EIN, FEC ID or government credential; without one, an individual can still run authorized ads under a verified name or their own legal name, labeled "About this ad." Idaho's thresholds start at $100, which is low but is also a few lines on a form. The disclosure rules were upheld in Citizens United and are explained in campaign finance lesson 05; whether the burden on small spenders is right is a fair argument for the Legislature. An individual who wants to speak without any of it can post unpaid, which Guidance 2026-5 says is generally not reportable.

If engagement is 66 percent bots on X, why be on X at all?

Pew's figure was for links to popular websites in 2017, not for all engagement, and it varied by site type. It does mean X metrics should be discounted and that follower growth there is not a reliable goal; M+R's nonprofits lost followers on X in 2025 and 31 percent planned to leave (M+R). Lesson 03 gives the decision table. The usual answer for an Idaho effort is to keep an X account for reporters and officials who still live there and to grow elsewhere.

Why trust growth benchmarks from brands and large nonprofits for a volunteer page?

You should not trust them as predictions; you should use them as the only published anchors and then replace them with your own numbers by week 13. The lesson's arithmetic shows what the anchors imply (13 a week for a 3,000-follower Instagram page) so that a plan promising 100 a week is recognizable as a fantasy before anyone commits to it. A small local page may beat the brand averages or trail them. Either way, after one quarter of honest logging it has a benchmark of its own.