Organic Versus Paid Social, Compared Honestly

The organic versus paid argument is usually conducted by people who make money from one of them. Here is the version with citations.

What each one actually is

Organic reach is distribution allocated by a ranking model, for free, based on predicted interest. Paid reach is distribution bought at auction, delivered to an audience you specify, whether or not a ranking model would have chosen you.

They are not two settings on the same dial. On LinkedIn, the help documentation states it directly: "Feed distribution isn't influenced by payments from third parties to LinkedIn, except for promoted (paid) content, which is clearly labeled." Ad spend does not raise your organic rank, and refusing to spend does not lower it.

Meta is equally clear about what boosting is. Its help documentation defines boosted posts as "ads you create from existing posts on your Facebook Page," and notes they are "still considered ads because they require a budget to be shared with a wider audience." Boosting is not a growth feature. It is the simplest possible ad, with fewer controls than a real campaign.

The thing paid genuinely gives you

Certainty of delivery. Organic reach is a prediction. Paid reach is a purchase. If you have an event on the 14th, paid is the only mechanism that guarantees anyone hears about it before the 14th.

Targeting you cannot get organically. Geography, life events, custom audiences from your own customer list, retargeting people who visited a page. Organic ranking will not respect your service radius.

Speed of learning. You can test four versions of an offer in a week for the cost of a decent dinner and find out which one people click. Organic testing takes months and confounds everything.

Reaching people who have already met you. Retargeting warm audiences is the least glamorous and most reliably useful thing a small budget does.

The thing organic gives you that paid cannot

Compounding. Nothing you buy accumulates. The moment the budget stops, the delivery stops. A YouTube video that answers a question, or a Pinterest Pin that Pinterest says can "gain engagement hours, days, months, or even years after it's first published," keeps working.

Evidence that a stranger chose you. A ranking model gave you a slot because it predicted a real person would want it. That is a signal about your business, and it is free market research.

Relationships that produce referrals. Nobody has ever referred a business to a friend because of an impression.

The uncomfortable research

Now the part almost no marketing article will tell a small business.

Measuring advertising returns is genuinely hard, and hardest at small scale. Randall Lewis and Justin Rao published a paper in the Quarterly Journal of Economics in 2015 analyzing "twenty-five large field experiments with major U.S. retailers and brokerages, most reaching millions of customers and collectively representing $2.8 million in digital advertising expenditure." Their finding: "the median confidence interval on return on investment is over 100 percentage points wide." Sales are so volatile relative to ad cost that "informative advertising experiments can easily require more than 10 million person-weeks, making experiments costly and potentially infeasible for many firms."

Read that again with your own budget in mind. Twenty-five well run experiments at large retailers could not pin down ROI to within 100 percentage points. Your $400 test cannot do better.

Some paid channels have been shown to buy customers who were already coming. Blake, Nosko and Tadelis ran large scale field experiments at eBay, published in Econometrica in 2015. Their result: "returns from paid search are a fraction of conventional non-experimental estimates. As an extreme case, we show that brand-keyword ads have no measurable short-term benefits." For non-brand keywords, "new and infrequent users are positively influenced by ads but ... more frequent users whose purchasing behavior is not influenced by ads account for most of the advertising expenses, resulting in average returns that are negative."

That was paid search, not paid social, and eBay is not your business. But the mechanism generalizes uncomfortably well: ads shown to people who were already going to buy from you look fantastic in a dashboard and change nothing.

The dashboards themselves are the problem. Gordon, Zettelmeyer, Bhargava and Chapsky used "data from 15 US advertising experiments at Facebook comprising 500 million user-experiment observations and 1.6 billion ad impressions" to compare experimental results with the observational methods advertisers normally use. Their conclusion, published in Marketing Science: "commonly used observational approaches based on the data usually available in the industry often fail to accurately measure the true effect of advertising."

Two of those four authors worked at Facebook. This is not an anti-advertising paper. It is a warning about the numbers in the interface.

So when is paid worth it

Paid is worth it when three things are true at once:

1. You know what a conversion is worth to you. Not "brand awareness." A booking, a quote request, an order with a margin you can state.
2. You can afford enough volume for the system to learn. Meta's documentation says ad sets "exit the learning phase as soon as they can deliver stably. This usually occurs after about 50 results in the week after the ad set's last significant edit." If a result costs you $20, that is roughly $1,000 in a week for one ad set. Below that, you are paying for a system that never gets out of exploration.
3. You have somewhere for the click to land that already converts. Paid traffic to a weak page is the fastest way to buy proof that your page is weak.

When it is money lit on fire

• Boosting a post because the app suggested it, with no offer and no destination.
• Paying to reach people who already follow you, who would have seen a good post anyway.
• Buying followers, likes, or "engagement." On TikTok, follower count is not even a ranking input, and Meta names engagement bait as a demotion category.
• Running an ad set at a budget too small for it to leave the learning phase, then judging it after four days.
• Spending on awareness when what you actually lack is a reason for anyone to care.

The honest recommendation for a small business

Use organic to find out what people respond to. Use paid to put more weight behind the two or three things that already worked, and to reach people organic cannot address, like a service area or a past customer list.

That order matters. Paid amplification of a message you have not tested is just a faster way to be ignored.

And the least expensive test available to any small business is still the oldest one: publish the thing, and find out whether people you respect find it useful. That is cheaper than an ad account and, at the sizes most businesses operate at, considerably more informative. It is also the whole design of a Ripple Room session.

When this was last confirmed

The Meta and LinkedIn documentation quoted here was checked in August 2026, with the boosted posts and learning phase pages read from archived snapshots dated May and August 2026 respectively. The three academic papers are peer reviewed and dated 2015, 2015 and 2019. Their findings are about measurement, which does not expire, but the platforms and ad products they studied have changed.