SEO vs Social Media: A Test, Not a Debate

We pulled the search results for seo vs social media from Ahrefs on 18 September 2026, ten positions deep. The number one result is a Reddit thread. Position six is Quora. Positions three, five, eight and nine are comparison articles that all open with "what is SEO" and close with "it depends on your goals".

Two of the ten slots on a page about whether social media matters are occupied by social platforms. That is the first useful thing this query teaches, and none of the articles ranking for it mentions it.

The second useful thing is that the debate itself is unanswerable. Nobody can tell you which channel is better without knowing whether anyone searches for what you sell, how long your content stays alive, and what your visitors do after they land. Those are three measurements, not three opinions, and you can run all of them this week on data you already have.

The debate is unanswerable, the question underneath is not

Search and social answer different demand, and which one your business has is a fact about your business, not a preference.

Search captures demand that already exists. Somebody types emergency plumber near me or best invoicing app for contractors because they have already decided they want the thing. Your job is to be there when they look.

Social creates demand that did not exist five seconds ago. Somebody scrolling a feed was not looking for you. Your job is to interrupt them well enough that they want something they had not thought about.

That distinction sounds academic until you apply it to a specific business:

The roofer who spends a year on Instagram is fighting to create demand that already exists on Google. The founder of a product nobody has a word for yet is fighting to capture demand that nobody is searching for. Both are working hard in the wrong place.

So the real question is not "which is better". It is "which kind of demand does my business run on", and that has a measurable answer.

Test 1: does search demand for what you sell exist at all

This is the check that can end the argument in twenty minutes, and it is the one nobody in the top ten runs.

If nobody searches for what you sell, search optimisation cannot work. Not because your writing is bad or your site is slow, but because there is no query to rank for. Quality is irrelevant to a keyword with zero volume.

There are two versions of the test, depending on whether you already have traffic.

If your site already gets impressions, Search Console has the answer. Pull ninety days of queries and count how many are not your brand name:

# Requires an OAuth access token with the webmasters.readonly scope.
curl -s -X POST \
  "https://www.googleapis.com/webmasters/v3/sites/sc-domain%3Ayourdomain.com/searchAnalytics/query" \
  -H "Authorization: Bearer $ACCESS_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "startDate": "2026-06-20",
    "endDate": "2026-09-18",
    "dimensions": ["query"],
    "rowLimit": 1000
  }'

Each row comes back with keys, clicks, impressions, ctr and position. What matters is not the totals. It is the shape of the list once your brand is removed:

// demand-test.js
// Feed it the rows array from the response above.
const BRAND = /yourbrand|your brand/i

const nonBrand = rows.filter(r => !BRAND.test(r.keys[0]))
const withRealDemand = nonBrand.filter(r => r.impressions >= 10)

console.log(`Non-brand queries:          ${nonBrand.length}`)
console.log(`With 10+ impressions:       ${withRealDemand.length}`)
console.log(`Non-brand impressions:      ${nonBrand.reduce((s, r) => s + r.impressions, 0)}`)

Three outcomes, three different decisions:

That third row hides two very different problems, and confusing them wastes quarters. A page that was never indexed produces the same zero as a query nobody types. Before concluding that your market does not search, confirm your pages are actually in the index: our guide to why a website does not show up on Google separates the four causes with Google's own status strings.

If you have no traffic yet, you cannot use your own data, so use volume estimates. Type the words a customer would use, not the words you use internally. A founder who calls the product a "revenue intelligence platform" will find no volume, while sales forecasting spreadsheet template has plenty. The vocabulary gap is the most common false negative in this test.

Free tools bound the answer well enough to decide. Our DIY SEO walkthrough covers the free keyword workflow in detail, and is SEO worth it turns the volume number into the only arithmetic that matters: clicks times conversion rate times value per customer.

A keyword with 30 searches a month is not automatically a failure, and one with 30,000 is not automatically a win. What decides is whether the searcher is close to buying.

Test 2: half-life, the number that separates the two channels

Ask any marketer why search compounds and social does not, and you get a metaphor. Measure it and you get a number.

A social post has a half-life measured in hours. Most of its reach arrives in the first day, the rest trickles in, and by week two it is functionally dead. A ranking page has the opposite shape: it earns little at first, climbs as it is crawled, evaluated and linked, and can be earning more in month twelve than in month two.

You do not have to take that on faith. Both platforms export daily numbers, so compute the decay yourself:

// half-life.js
// series: array of { date: 'YYYY-MM-DD', value: Number }, oldest first.
// value = daily clicks (Search Console) or daily reach (your social platform).

function cumulativeShareByDay(series, day) {
  const total = series.reduce((s, p) => s + p.value, 0)
  if (total === 0) return 0
  const upTo = series.slice(0, day).reduce((s, p) => s + p.value, 0)
  return upTo / total
}

function halfLifeDays(series) {
  const total = series.reduce((s, p) => s + p.value, 0)
  let running = 0
  for (let i = 0; i < series.length; i++) {
    running += series[i].value
    if (running >= total / 2) return i + 1
  }
  return series.length
}

console.log('Share delivered in first 48h:', cumulativeShareByDay(series, 2).toFixed(2))
console.log('Days to half of all delivery:', halfLifeDays(series))

Run it on one social post and one blog page, both at least ninety days old. The two numbers will not be close, and the gap is the entire strategic difference between the channels.

Two honest caveats about this test.

The first is that a page with a short half-life is not proving that search does not compound. It is proving that this page does not rank, which is a different problem with a different fix. A page stuck on page four has no curve to measure.

The second is that the comparison flatters search if you count only production time. Search work has recurring costs that never appear in a "write it once" story: internal linking upkeep, pruning, and rewriting titles that earn impressions but no clicks. Our content audit method is what that upkeep looks like in practice, and how much SEO costs prices it.

If you want the timeline version of this measurement rather than the decay version, how long SEO takes dates each gate a page passes through, from first crawl to first click.

Test 3: which channel produces behaviour, not just traffic

The third test is the one that changes minds, because it regularly contradicts the first two.

Traffic is not the product. Signups, bookings, carts and calls are. A channel that sends a tenth of the visitors and a third of the customers is the better channel, and you only find that out by separating the two sources and comparing what happens after the click.

Search traffic is easy to isolate: Search Console reports it, and your analytics reports the organic referrer. Social traffic is harder, and almost everyone undercounts it for three reasons:

• In-app browsers strip referrers. A tap from a native app often arrives as direct traffic.
• Dark social. A link pasted into a group chat or a DM carries no source at all.
• Link-in-bio hops. The intermediate page becomes the referrer, hiding the platform.

The fix is boring and it works: tag every social link with UTM parameters, every time, no exceptions.

https://yourdomain.com/pricing?utm_source=instagram&utm_medium=social&utm_campaign=bio-link
https://yourdomain.com/blog/your-post?utm_source=reddit&utm_medium=social&utm_campaign=sept-thread

Then compare the two cohorts on the same event, not on visits. Conversion rate, revenue per session, or qualified calls booked. If your social visitors convert at a third of the rate of your search visitors, that is not a reason to abandon social. It is a reason to stop comparing the channels on traffic, which is the number both platforms are happiest to show you.

We wrote the two-source measurement setup in detail for the paid comparison, and it transfers unchanged: SEO vs Google Ads has the section on measuring two channels in one place without letting either platform grade its own homework.

The overlap: social platforms are search results now

Here is where the "versus" framing breaks down entirely, and the evidence is the page you would land on if you ran this search yourself.

In the results we measured on 18 September 2026, a Reddit thread holds position one with a domain rating of 95, and Quora holds position six. Neither is a website in the sense the debate assumes. Both are social platforms appearing as organic search results.

This is not specific to marketing queries. Forum threads have been taking visible share of informational results for the last two years, which is why "post where people already are" and "rank in search" have stopped being opposite strategies for some queries.

Two changes made that concrete:

• Instagram opened public content to indexing on 10 July 2025. Its help documentation sets the conditions: public professional accounts, account holder over 18, and posts from 1 January 2020 onward. The setting is on by default for eligible accounts and can be switched off in privacy settings.
• Google and Reddit signed a content licensing agreement in February 2024, reported at around $60 million a year. As of publication, press reports say the deal is near expiry and its renewal is uncertain, so treat the commercial arrangement as temporary and the underlying behaviour as durable.

The practical consequence is a workflow, not a slogan. Before you decide a query belongs to social or to search, look at what already ranks for it. If forum threads hold several slots, a genuinely useful answer posted in the right community can earn search visibility that your own domain would need a year to earn. If the results are all vendor pages and documentation, posting in a community will not touch that query.

The same logic applies to how AI assistants answer. They cite what they can retrieve, and social threads are retrievable. Our work on getting cited by ChatGPT and on LLM visibility covers which surfaces get quoted and how to check whether yours is one of them.

What Google actually says about social signals

Google's own documentation is narrower than either camp claims, and worth quoting rather than paraphrasing.

The SEO Starter Guide treats social media as a discovery channel, not a ranking lever. It lists social media promotion among the ways to spread the word about a site, and frames the benefit as speed of discovery: "Effectively promoting your new content will lead to faster discovery by those who are interested in the same subject, and also by search engines."

The mechanism it names is links, not likes. The same guide states that "the vast majority of the new pages Google finds every day are through links, making links a crucial resource you need to consider to help your pages be discovered by Google."

Read those two sentences together and the honest version of the relationship falls out. Likes, shares and follower counts are not documented ranking inputs. A post that travels far can still help you rank, because people who see it may link to you, mention you, or search for your brand later. The social activity is upstream of the thing that counts, which is why the correlation is real and the causation is indirect.

Google is equally blunt about outcomes, and it is the sentence every channel comparison should carry:

No one can guarantee a #1 ranking on Google.

That is verbatim from Google's guidance on hiring an SEO. Results vary by site, by market and by competition, and any comparison that hands you a confident forecast is selling something. The starter guide says the same thing about timing: some changes take effect in a few hours, others take several months.

Allocating five hours a week, honestly

Most readers of this question are not choosing a budget. They are choosing what to do with the only free evening they have, which makes "do both" an expensive answer dressed as a safe one.

Split across two channels, five hours a week buys a thin presence in each. Concentrated, it buys one article a week that can rank, or three posts a week that can build an audience. The three tests above tell you which concentration fits:

Two cases where one channel is simply wrong for a while.

If your product solves a problem people do not know they have, search cannot find you a buyer. Nobody types a query for a solution they cannot name. Create the demand on social, and expect search demand to appear later, first as brand searches.

If your customers are in acute need, social is a poor fit. A burst pipe does not get solved by a well-crafted Reel. That buyer is on Google right now, and the only question is whether you are there.

Whatever the split, the channel that earns the next hour is the one that the tests say is working, not the one that feels more productive. Publishing is satisfying. Measuring is not. The founders who get this right check the numbers on a schedule and move hours between channels on evidence, which is the same discipline that separates content marketing that earns revenue from content marketing that fills a calendar.

What to do this week

Four steps, in order, all finishable in an afternoon:

1. Run the demand test. Pull ninety days of Search Console queries, strip your brand, count what is left. If you have no data, check volume on the five phrases a customer would actually type.
2. Compute one half-life on each side. One social post, one page, both at least ninety days old. Write both numbers down.
3. Tag every social link with UTM parameters and give it two weeks. Then compare conversion rate, not sessions.
4. Read the results page for your main query. If forums hold several slots, a useful answer in the right community is a search play. If they do not, it is not.

Then pick the concentration those four answers point to, and revisit in a quarter. The answer changes as your site earns authority and your category earns searchers, which is why "SEO or social media" is a question you re-answer rather than settle. For a solo creator whose audience lives on one platform, the starting point often looks different from a SaaS founder's: our note on SEO for musicians is the version of this trade-off for people whose discovery already happens in a feed.

FAQ

Is SEO more important than social media?

Neither is universally more important. The determining fact is whether demand for what you sell already exists as search queries. If people type what you sell, search captures buyers closer to the decision and keeps doing it after you stop working. If they do not, social is the only channel that can create that demand, and search becomes useful later, once people know the words for your product.

Is SEO being phased out?

No, though the surface is shifting. AI answers and richer results absorb clicks that used to land on websites, which changes what a given ranking is worth. The underlying activity is unchanged: people still ask machines for things, and those machines still retrieve pages. The practical adjustment is to measure impressions and citations alongside clicks, and to write pages worth quoting rather than pages that rank and say nothing. Our work on answer engine optimization covers what the shift changes and what it does not.

Can you use SEO for social media?

Yes, and for many queries this is the highest-return version of the question. Social platforms run their own search, and their public content increasingly appears in Google's results too. Instagram opened public professional accounts to search engine indexing in July 2025, and forum threads rank for a large share of informational queries. Write post captions and thread titles with the phrasing people search, not just the phrasing that performs in a feed.

What are the 4 types of SEO?

The common taxonomy is on-page (content, titles, structure), off-page (links and mentions from other sites), technical (crawling, indexing, speed, structured data) and local (Business Profile, maps, location signals). It is a convenient way to divide work rather than a distinction Google makes. Google's documentation talks about crawling, indexing, serving and helpful content, which maps only loosely onto those four buckets.

Does social media help SEO rankings?

Indirectly, through links and discovery rather than through engagement counts. Google's starter guide lists social promotion as a way to get content found faster, and names links as how most new pages are discovered. Likes and shares are not documented ranking inputs. A post that reaches the right person and earns a link, a mention or a later brand search does help, which is a slower and less predictable mechanism than the shortcut people hope for.

How do I know if my channel choice is working?

Set one event as the measure, such as signups or booked calls, and attribute it by source: Search Console plus organic referrer for search, UTM tags for social. Compare cohorts on conversion rate rather than on sessions, and give each channel at least a quarter before judging. Results vary between sites and markets, so your own numbers are the only ones that settle the question for your business.