Somewhere past the first few thousand in daily spend, every ecommerce account meets the same wall. Retargeting still converts. The creator testimonials still work on people who already know the category. Then prospecting budgets go up, the audience gets colder, and the blended numbers sag. The usual diagnosis is creative fatigue, so everyone makes more ads. But if the new ads are the same testimonials pointed at colder traffic, volume fixes nothing, because the people now seeing them don't yet believe they have the problem the ad is solving. That moment is what VSL ads, podcast-style ads and street interviews are for.
Eugene Schwartz described this wall in 1966. Breakthrough Advertising splits every market into five states of awareness, from buyers who know your product and are waiting for a reason to act, down to people who don't know they have the problem at all. At each state, Schwartz argued, the prospect is asking a different question, and marketing that answers the wrong one gets ignored. A testimonial answers "does this actually work?". Cold traffic is asking "why should I care?", or more often, nothing at all.
This piece covers the three formats built for the bottom of that ladder: what each one is, why it earns attention from people who weren't looking, roughly what it takes to produce, and a sensible way to test your first one. We've mapped which UGC script format fits each awareness stage before; this goes deeper on the three that do the hardest job.
Key takeaway
Standard UGC testimonials assume the viewer already understands the problem. VSL ads, podcast-style ads and street interviews are built to create that understanding first. Match the format to the awareness level of the traffic you're buying, not to what worked on retargeting.
Why your best ads die on cold traffic
Warm audiences do half the ad's work. Someone who has googled the category, follows a competitor or sat through your organic content brings context with them, so a 30-second testimonial lands. Push spend past those pockets and Meta starts buying you people with no context at all. The ad is the same. The question it answers stopped being asked.
You can see the ladder in the formats themselves. Testimonials and demo UGC serve solution-aware and product-aware viewers. Educational and entertainment-shaped formats serve problem-aware and unaware viewers. Our catalogue of 18 UGC ad formats covers the full menu; the three below are the ones we'd point at genuinely cold traffic, because each one manufactures the context the viewer doesn't have. One buys time, one borrows trust, one shows proof that visibly wasn't staged.
Cold traffic isn't rejecting your product. It's rejecting an ad that assumes an interest nobody gave it.
What is a VSL in ecommerce?
A VSL (video sales letter) is a scripted video that pitches one offer and asks the viewer to take one action. Vidico's 2026 guide defines the shape: hook, problem, solution, proof, offer, call to action, in that order, running anywhere from under a minute for a social cut to about 30 minutes for a detailed offer. The same guide notes that US digital video ad spend is set to pass $80 billion in 2026, up 11% year on year, citing IAB figures. Ecommerce VSLs live at the short end: a 60-second to 3-minute cut runs as the ad, and a longer version carries the landing page.
The format exists because cold traffic needs the one thing a testimonial can't give it: time. Thirty seconds is not enough to convince someone they have a problem, explain why what they've tried hasn't fixed it, and introduce a mechanism they've never heard of. Ninety seconds, scripted tightly, is. That's why VSLs sit behind so many of the supplement, device and telehealth offers that scale hardest on cold traffic, and why the format keeps outliving its own obituaries.
Here's the arc we'd script for a 90-second ecommerce cut:
| Beat | Timing | What it does |
|---|---|---|
| Hook | 0 to 3s | Name the symptom, not the product. The viewer should recognise themselves before they see a brand. |
| Problem | 3 to 25s | Make the problem specific and costly. Generic pain reads as advertising; detail reads as recognition. |
| Mechanism | 25 to 45s | Why the usual fixes fail, and the one idea that makes your product different. This is the part a testimonial skips. |
| Proof | 45 to 70s | Demo, numbers, a face saying it worked. Stack two or three kinds of proof, fast. |
| Offer and CTA | 70 to 90s | One offer, one action. Price anchoring and the guarantee live here, not in the hook. |
Production is cheaper than it looks, because the script is the product. Write it and table-read it before anyone films anything. Shoot modular (the problem section, the mechanism, each proof element as separate blocks) so the edit can produce a 60-second cut, a 90-second cut and a lander version from one shoot. The expensive mistake is filming a beautiful three-minute film first and discovering the hook doesn't hold.
Why podcast-style ads hook cold viewers
A podcast-style ad is a video ad shot to look like a podcast clip: two people across a table with microphones in frame, cut like a moment from a real show. The feed reads it as a podcast clip before it reads it as a placement, which buys the ad the seconds a cold viewer would never grant an obvious pitch.
Influee, a creator marketplace that sells the format (so weigh the source accordingly), reports hook rates above 40% across 12+ ad accounts running it internally, and recommends 30 to 60 second cuts for paid social. The same page cites Edison Research's Infinite Dial survey: 42% of Americans aged 12 and over listened to a podcast in the last month. The format borrows a listening habit the audience already has.
It suits cold traffic for a quieter reason too. A conversation can raise a problem the way a friend would, as an opinion rather than a claim. "I didn't realise how much caffeine was wrecking my sleep until..." lands differently across a microphone than straight to camera. The viewer is overhearing rather than being sold to, and overheard opinions meet less resistance.
Of the three formats this is the cheapest to test and the easiest to fake badly. If the room, the mics and the pacing don't read as a real show, you've made an expensive talking head. Use creators who can hold an unscripted conversation, give them talking points instead of lines, and spend the edit budget on the first 20 seconds, because that's where the format's trust either holds or collapses.
Why street interviews sell products nobody searches for
The street interview is the oldest format on this list and the newest performance channel. A host with a microphone asks strangers a question, the product appears mid-conversation, and the reactions are real because nobody handed the stranger a script. For an unaware viewer, that is the whole pitch: a person with no stake in the outcome reacting to a product they'd never heard of either.
It has quietly become an industry. StreetTalk, founded in 2024, works with DTC brands including PrizePicks, Ridge, Grüns, Brez and Dr. Squatch, and signed roughly 80 brands in its first summer. Its typical ad fits a first reaction, a live product experience and a testimonial inside 45 seconds, often tested organically before any paid spend goes behind it. Street Poller Media runs the same play at volume; its founder estimates he can field 150 interviewers within 24 hours and deliver 1,000 pieces of content in two weeks. Brands are buying a steady feed of unscripted proof, not one viral video.
The format rewards products that demo instantly, where someone can taste it or feel the difference in seconds. It punishes anything that needs a minute of context, because a stranger on a pavement can't supply context they don't have. It also carries the most production friction of the three. You need a host who can talk to strangers without making it weird, releases signed for every usable face, and the stomach for a high miss rate: most interviews are unusable, and the edit is built from the handful of real moments. That is also exactly why it converts. The moments you keep are visibly not acting.
Which format for which job
The three formats aren't interchangeable. They enter at different awareness levels and fail in different ways.
| Format | Awareness level it suits | Typical ad length | Production lift | Pick it when |
|---|---|---|---|---|
| Standard UGC testimonial | Solution aware and product aware | 15 to 45 seconds | Low | The viewer already knows the category and needs belief, not context |
| VSL ad | Unaware and problem aware, considered offers | 60 seconds to 3 minutes, longer on the landing page | Medium, and script-heavy | The product needs explaining, has a new mechanism, or carries a price that needs justifying |
| Podcast-style ad | Problem aware and solution aware, cheap reach on unaware | 30 to 60 seconds | Low to medium | You want a low-cost cold-traffic test and have a creator who can hold a real conversation |
| Street interview | Unaware and problem aware | 30 to 60 seconds | High: host, locations, releases, a high miss rate | The product demos in seconds and third-party reactions are the strongest proof you have |
How to test your first one without blowing the budget
Diagnose before you produce. If prospecting CPA is healthy and the problem is retargeting, none of this applies; your issue is further down the funnel. These formats earn their keep when cold prospecting is the thing that won't scale.
- Pick one format, not three. Use the table above. Considered product with a mechanism, go VSL. Instant demo, go street interview. Neither, start podcast-style, because it's the cheapest read.
- Make three hooks on one body. Same core video, three different opening lines or moments. Cold formats live or die in the first three seconds, and you want a verdict on the opening before you judge the concept. Our breakdown of UGC hooks that stop the scroll applies here unchanged.
- Run it cold only. Exclude site visitors, buyers and engaged audiences, otherwise warm pockets flatter the result and you learn nothing.
- Read hook rate first, CPA second. The format's job is to earn attention from people who weren't looking. If hook rate on cold traffic beats your current prospecting ads, the format is working; then give CPA enough spend to reach a verdict before you judge it.
- Feed the winner back into the system. A 90-second VSL cut that holds becomes the landing page video. A podcast clip that hooks becomes a series with the same pair. This is the loop our piece on building a creative system that drives growth describes.
Our view, for what it's worth after 5,000+ ads made: when cold prospecting stalls, a format change moves the numbers more often than another round of polish on the format you already run. The eleventh testimonial variant is rarely the answer. A format that builds context for the viewer usually is.
Where Spark fits
Full disclosure: Spark UGC is our studio. We plan creative against awareness stages, produce with vetted creators (and AI where it genuinely helps), and iterate on ad-level results, with brief to live ads taking about 21 days. If the account has hit the cold-traffic wall and you want the format strategy and the production handled by one team, how it works shows the process. If what you actually need is a five-city street shoot next week, a specialist shop like the ones above is the better call, and we'll say so.
VSL, podcast-style and street interview FAQ
What is a VSL in ecommerce advertising?
A VSL (video sales letter) is a scripted video that pitches one offer and asks the viewer to take one action. In ecommerce it usually runs as a 60-second to 3-minute ad on Meta or TikTok, with a longer version on the landing page, and follows a fixed arc: hook, problem, mechanism, proof, offer, call to action. It suits cold traffic because it has time to build the problem before presenting the product.
How long should a VSL ad be?
A full VSL can run from under a minute to about 30 minutes for a detailed offer. For paid social, test 60 to 90 second cuts first: long enough to build a problem and a mechanism, short enough to hold a feed audience. Let the landing page version carry the full argument.
What are podcast-style ads?
Podcast-style ads are video ads shot to look like a podcast clip: two people in conversation, microphones in frame, natural pacing, usually 30 to 60 seconds for paid social. They work on cold audiences because the feed reads them as content before advertising. Influee, a marketplace selling the format, reports hook rates above 40% across 12+ ad accounts running it.
Do street interview ads work for small brands?
Yes, if the product demos quickly and you plan for the miss rate. Start with one host, one location and signed releases for every usable face, then test the clips organically before putting paid spend behind the strongest reactions. The format's proof value comes from reactions that are visibly unscripted, which no budget can fake.
Cold audiences are where the next tier of growth lives, and they can't be reached with ads built for people who were already halfway convinced. If prospecting has stalled, tell us what you're running and where it stops working, or dig further into testing and formats in the resources library.