Live shopping sessions exploded. Live shopping value did not. - Feature 43 | Social Intelligence

Three figures, and the relationship between them is the point.

Livestream commerce sessions surged 340%. US live commerce grew 42% year over year. And live currently accounts for roughly 14% of US social commerce GMV.¹

Figure 01

Live commerce, three measures

Session growth, market growth and current share.

Bar chart showing livestream commerce sessions up 340 per cent, US live commerce growth of 42 per cent, and live at 14 per cent of US social commerce GMV.

These three bars measure different things and are not comparable to each other. One is session count growth on one platform, one is market value growth, one is a share of merchandise value. They are shown together to describe a format's position - high activity growth, solid value growth, small share.

Sessions up 340%, value up 42%, share at 14%. Supply is arriving far faster than demand, which is what the early part of a format's curve looks like - and also what a bubble looks like.

The gap between 340 and 42

That divergence is the most informative thing here and it goes unremarked in the coverage.

If sessions grew 340% while the value of live commerce grew 42%, then the average session is producing substantially less than it did a year ago. Far more people are going live; the money is not keeping pace.

Two readings, and they have opposite implications.

Figure 02

Two readings of the same divergence

Sessions growing eight times faster than value.

Healthy

Supply arriving ahead of demand

New entrants start smallA million new sellers running their first streams will average near zero, dragging the mean down while the total grows.

The format is being learnedLive selling is a skill. Early sessions from new participants underperform by construction.

Share is the metric to watch14% rising is the signal; average session value is noise during onboarding.

Unhealthy

Volume chasing a fixed audience

Viewer time is finiteEight times the sessions competing for the same hours means most streams find nobody.

Platform incentives inflate supplySeller onboarding programmes add sessions regardless of whether demand exists for them.

Share may stallIf 14% plateaus while sessions keep climbing, the format has found its ceiling.

Both readings are consistent with the three published figures and we cannot distinguish them. The deciding data - average revenue per session over time - is not published, and it is the single number that would settle which curve this is.

We lean slightly toward the first reading, because the 1.2 million new merchants documented in Feature 40 would mechanically depress average session value during onboarding. But that is a lean, not a finding, and anyone building a business on live commerce should want the session-value series before committing.

Why live converts

Set aside the growth question and the format's mechanics are worth understanding on their own, because they are unusual.

Reported effects are large: live shopping drives 10 to 15 times more engagement, and brands running weekly streams see 3 to 5 times higher conversion than those relying on feed posts alone.¹

Those are vendor-published multiples of exactly the kind Feature 12 of the Content desk taught us to distrust - no baseline, no sample, no definition of engagement. We report them as claims rather than findings.

The mechanism, though, is legible without the numbers. Live selling reintroduces three things that e-commerce removed:

Answered objections. A viewer asks whether it fits, how it washes, whether it works on their skin - and gets an answer in seconds, from a person handling the product. Every unanswered objection in conventional e-commerce is an abandonment.

Scarcity that is actually true. A limited quantity in a limited window is the oldest selling mechanic there is, and in a live stream it is verifiable rather than asserted. Countdown timers on product pages are not believed; a stock counter falling in real time is.

Demonstration under conditions the buyer sets. Not a produced video approved by a brand team, but a product being used in response to a stranger's request. This is the property Feature 13 of the Ads desk found consumers respond to - and it cannot be produced in advance.

None of these is new. All three are what a good market stall has always done, and what a decade of conversion-rate optimisation could not replicate through static pages.

Why the West lagged

Live commerce has been mature in China for years and slow everywhere else, and Feature 08 of the Ads desk explains why better than any cultural argument.

Live selling requires content and transaction on one surface. A stream that sends viewers to another site to buy loses the scarcity, the immediacy and the answered objection - the entire mechanism. Chinese platforms built commerce into content from the start. Western platforms built content first and are now retrofitting commerce, which is the same lag Feature 40 measures in the conversion spread.

The 14% share is therefore not evidence that Western consumers dislike live selling. It is evidence that the infrastructure only recently became capable of supporting it, and the 340% session growth is what happens immediately afterwards.

Whose interest this serves

The engagement and conversion multiples in this feature - 10 to 15×, 3 to 5× - are published by agencies and platform partners selling live commerce services, with no stated baseline or method. We report them as claims and grade them accordingly. The growth figures come from the same sources and carry the same interest. Marketing Legendary operates a social practice and a conclusion that live commerce is early rather than saturated is a conclusion that there is work to sell.

What to do about it

Ask for average revenue per session before you commit. It is the number that distinguishes the two readings in Figure 02, and any partner proposing a live programme should be able to produce it for comparable sellers.

Treat weekly cadence as the threshold, not the aspiration. The conversion claim attaches specifically to brands running weekly streams. An occasional stream is a different product with no evidence behind it.

Staff it for objection handling, not for presenting. The mechanism is answering questions in real time. A polished presenter who cannot answer a fit question is worse than an unpolished operator who can.

Do not run it where checkout is a handoff. Feature 40's architecture argument applies with more force here. Live selling on a surface that sends people elsewhere to buy discards the reason the format works.

Discount the multiples, keep the mechanism. 10 to 15× engagement is an unverified vendor claim. Answered objections, real scarcity and responsive demonstration are structural properties you can reason about without trusting anyone's number.

Figure 03

Live commerce, 2026

Growth, share and claimed effects.

Measure	Value	Grade
Growth and share
Livestream commerce session growth	+340%	Reported
US live commerce value growth	+42%	Reported
Live share of US social commerce GMV	~14%	Reported
New merchants onboarded	1.2m+	Feature 40
Claimed effects
Engagement multiple	10–15×	Unverified, no baseline
Conversion multiple, weekly streamers	3–5×	Unverified, no baseline
Not established
Average revenue per session, over time	-	Not published - the deciding number
Whether 14% is rising or plateauing	-	Single observation
Viewer time available to the format	-	No data

The three growth and share figures measure different things - sessions, market value and share of GMV - and are not comparable to one another.

How we did this

Where this comes from

From a company that sells into this market: live and social commerce benchmark compilations from agencies and platform partners. No Tier 1: no platform disclosure or dataset obtained.

Not comparable

The 340%, 42% and 14% figures measure different quantities. We compute nothing between them beyond noting the direction of the session-versus-value divergence.

Graded down

The 10–15× and 3–5× multiples are vendor claims with no stated baseline, sample or definition. We report them as claims, not findings, per the standard set in Feature 12 of the Content desk.

What's ours, not the source's

The two readings in Figure 02, the three-mechanism explanation and the infrastructure-lag argument are ours.

What this doesn't prove

Whether live commerce is early or saturating. Both readings fit the data. The deciding series is not published.
That the engagement and conversion multiples are real. Unverified vendor claims with no baseline. We report the mechanism instead.
That the three mechanisms cause the conversion. They are structural properties of the format, not tested variables.
That infrastructure explains the Western lag. It is our argument, consistent with Feature 40's conversion spread. Cultural and regulatory differences are untested alternatives.
Whether the 340% session growth continues or reverses. Single observation, driven partly by onboarding programmes that will not repeat.
Anything about profitability. No source reports the cost of running live commerce against what it returns.

Sources for this feature

Live and social commerce data, 2026. shortformnation.com, shortsintel.com, digitalapplied.com From a company that sells into this market - agency and platform-partner compilations
Features 40 and 42 of this desk, Features 08 and 13 of the Ads desk, Feature 12 of the Content desk. Another feature in this edition