2026-09-10 · Austin

The honest ROI of AI video for a small business — a real number

Every AI content tool on the internet right now is selling the same thing: post more, grow faster, make money in your sleep. Nobody wants to publish the boring math on whether it actually pays off.

I ran the experiment. Three small businesses I know — a mobile coffee cart, an antique reseller, and a general contractor — agreed to post one AI-generated short-form video per day for 30 days on TikTok and Instagram, using our tool. I paid attention to what happened. Here are the actual numbers, with no rounding up.

Setup

3 small businesses, 3 different verticals.
30 days of posting: one 15-30 second AI-generated vertical video per day per platform (TikTok + Instagram Reels), so 60 posts per business, 180 posts total.
Each business already had a small existing audience (< 1,000 followers).
Each used a mix of AI-generated video from a text prompt, plus stitched short clips from their own footage.
Total cost per business for the month, all-in (tool subscription + provider API charges + their time editing prompts): roughly $85 to $110.

What happened in aggregate

Followers gained across all 3 accounts: 4,847 new followers over the 30 days (average ~1,600 per business).

Total views across all posts: 891,000. Median video: about 3,200 views. Top video: 340,000 views (an antique-flip one that hit TikTok's algorithm on day 12). Bottom 25% of videos: under 400 views each.

Actual attributable revenue:

Coffee cart: 42 new customers who mentioned finding them on social. At average $8/order, roughly $340 in first-visit revenue, plus assumed repeat purchases.
Antique reseller: 6 direct sales sourced from Instagram DMs about specific pieces they posted. Total $1,180.
Contractor: 3 leads that turned into estimate visits. 1 closed into a $4,400 deck project.

Add it up: about $5,920 in first-touch attributable revenue against $285 in tool costs. That's a 20:1 return for the whole cohort.

But that's the wrong number to look at

If I just told you "20x ROI" that would be misleading. Here's what's underneath it:

The contractor's $4,400 deck job carried the entire cohort

Without the deck project the ROI is more like 5-7x. Which is still positive but nowhere near a "quit your day job" outcome. Big-ticket service businesses can get outsized returns from social because one closed job pays for months of content. Retail and impulse-buy businesses need way more volume for the numbers to work.

The follower gains are real but half of them are useless

Roughly half the followers each account gained were what I'd call "AI content watchers" — people who follow lots of accounts, engage with the algorithm, but never buy anything from anyone. The other half were locals and enthusiasts. The half that matters is smaller than the follower count suggests. If you're being honest with yourself, divide any follower gain from AI content by two before you calculate what it's worth.

The videos that worked were the ones with real footage in them

Pure text-to-AI videos got the lowest engagement across all three accounts. The videos that mixed AI-generated B-roll with actual clips of the person, product, or workspace consistently outperformed. This surprised me a little but shouldn't have. People buy from people. AI removes friction; it doesn't remove trust.

The time cost is real even with AI

The tool cuts editing time from ~2 hours per video to ~15 minutes. But that 15 minutes still has to happen every day. Every business owner in this cohort said the hardest part wasn't the videos themselves, it was doing it consistently for 30 straight days. Two of the three admitted they missed 2-4 days somewhere in the middle.

The actual honest ROI framework

Here's the way to think about it if you're a small business considering this:

Multiply the following:

Your average customer lifetime value (be conservative — first purchase minus cost of goods).
Your typical conversion rate from social traffic (for most local businesses this is 0.5% to 2%).
Roughly 3,000 views per post (realistic median with consistent posting).
30 posts per month.

For a coffee shop: $8 × 0.02 × 3,000 × 30 = $14,400 in top-of-funnel value per month if everything works. Realistically 15-25% of that: $2K-$3.5K/month.

For a contractor: $2,000 average project × 0.01 × 3,000 × 30 = $180,000 in top-of-funnel value per month if everything works. Realistically 5-10% of that: $9K-$18K/month with one or two closed jobs.

For an antique reseller: $80 × 0.015 × 3,000 × 30 = $10,800 top-of-funnel per month, realistically $1.5K-$3K/month in real sales.

Now subtract:

Tool cost: $19-$50/month
API/generation cost above the tool's cap: probably $30-$80/month if you post daily
Your time: 15 min/day × 30 = 7.5 hours/month. Value that however you value your hourly time.

For most small businesses, the math is positive but not obscene. It's a lever, not a lottery ticket.

When it does not work

Being straight: AI content is a bad investment for you if:

You don't have consistent posting discipline. The tool doesn't post for you. It makes the posts. You still have to hit publish. If you can't commit to 3-5 posts a week for at least 60 days, don't bother.
Your business is invisible on video. SaaS-to-SaaS enterprise sales, most B2B, high-touch consultancies — social video is not your channel. Don't force it.
You're expecting instant results. Every account in this experiment saw nothing for the first 8-12 days. If you'd measured on day 10 you'd have called it a failure. The algorithm needs a data set before it decides who to show your stuff to.
You're chasing follower count instead of business outcomes. If your revenue model doesn't convert social attention into money, more attention won't fix that.

What we recommend

If you're a small business considering AI content tools, our honest advice:

Start with a 30-day commitment. Not a week. Not "I'll try it." Thirty consecutive posting days. If you can't do that, save the money.
Mix AI-generated content with real footage of you, your team, your product. 70% AI / 30% real is a good starting ratio for a business, not 100% AI.
Measure at day 60, not day 15. The compounding effect of consistent posting is real but slow.
Pick one platform to be great at before adding a second. Most small businesses should start with either TikTok or Instagram based on where their customers actually live.
Track conversions, not followers. Followers are a lagging indicator of a lagging indicator.

That's the honest picture. AI content works. It doesn't 10x your business, but it can 2-4x the return on the time you're already spending on marketing, which for most small businesses is worth doing.

If you want to try it for yourself, Poster gets you one free video and a 7-day trial of everything else. No card required. Enough to test the math for your own business without committing to another subscription.

Written in Palmerton, PA. Not Silicon Valley. Not a growth-hacker Slack. We build the way carpenters build — measured twice, cut once.