3 weeks in, 2 free trials from strangers — what actually surprised me
Launched verism.app about 3 weeks ago. Two products — a career simulator that puts you through a full day in 9 different careers (will add more) with real decisions and a behavioral debrief, and a situation simulator that's voice based where you practice salary negotiations and performance reviews against an AI that responds like a real manager would.
sharing what surprised me because some of it I genuinely didn't see coming.
Carousels on TikTok dramatically outperform videos. My screen recording videos failed in the first second, people swiped immediately. Carousels with a strong opening line held attention for 10 to 20 seconds. didn't expect the format to matter that much.
Tried three different promotion types — get views, get likes, and get website clicks. Getting views was basically useless, people watched for 1.5 seconds and moved on. Get website clicks drove real traffic but converted at 0.1%. Get likes drove the most saves and engagement but people are saving and not clicking through to the site. around $100 spent total, 2 free trials from strangers, 0 paying users.
The saves are actually what confuses me most. Some posts have really strong save rates which suggests people find it interesting enough to come back to later, but the conversion to actually trying the product hasn't happened. Not sure if that's a trust problem, a friction problem, or just cold traffic that's never going to convert.
What I'm still figuring out is how much of the bad conversion rate is a cold traffic problem versus something actually wrong with the product or homepage. I think it's mostly the traffic but I don't have enough real strangers through yet to know for sure.
Also genuinely unsure whether the two products belong together or whether having both is confusing to a cold visitor. They share a thesis — experience high stakes moments before the real ones — but the audiences are a bit different.
open to any thoughts. especially curious if anyone has made premium B2C work without being the face of it and what actually moved the needle.
verism.app if you want to look
the save-vs-click gap usually isnt a traffic problem, its a positioning gap on the LP. carousels work bcs the first frame does positioning in 2 sec — your LP prob doesnt repeat that same first-frame promise within 3 sec of landing. premium B2C without a face is doable but needs a sharp "category of one" framing imo. ngl 0.1% click conversion on cold tiktok sounds normal — the bigger Q is what % of the 2 trialers told u in plain words why they trialed. if u havent asked them yet, thats where the next 5 hrs of work should go before scaling any channel.
Worth pushing on what Warren455 touched but didn't take far enough — the save behavior isn't ambiguous, it's diagnostic. Cold TikTok audiences save self-improvement content the way they save recipes they'll never cook. The save is the dopamine hit; trying the product is the work they're avoiding. You're not getting bad conversion — you're getting the exact conversion that audience produces on that promise.
The harder question hiding in your post: a free trial is the wrong mechanic for cold B2C self-improvement. Trials work when someone already has the problem and is actively shopping. For your audience, the friction stack is brutal — sign up, remember to use it, complete a session, decide it's worth paying, come back to convert. That's 5 commitments from someone who saved a TikTok 3 days ago.
What usually works better: a $5-9 one-time purchase for a single simulation with a money-back guarantee. Lower the commitment, raise the trust signal. People pay $7 for curiosity they'd never start a trial for.
The saves tell you something useful actually. They mean people find the concept interesting enough to return to. The gap between a save and a signup is almost always one of two things: either they do not fully understand what they are getting into before they commit, or the moment they land on the site they cannot picture themselves as the person who would use it.
The cold traffic question you are asking is real but it cannot be answered from analytics. The only way to know whether it is a traffic problem or a product problem is to put five people who match your target audience through the product and watch where they hesitate. Not ask them afterwards. Watch it happen.
My guess based on what you have described is that the two products being side by side is creating a decision moment for a cold visitor that should not exist yet. They arrive, see two things, and have to figure out which one is for them before they have even decided if either is for them. That extra cognitive step is probably costing you.
One product, one person, one problem. See if that changes the conversion rate before you spend more on traffic.
If you want to run a structured feedback cycle with real users through EarlyLoop, get in touch. earlyloop.uk. This is exactly the kind of question five real people through the product would answer in a week.
The “save but no signup” thing is interesting honestly.
I see this a lot with AI tools lately.
People save because the idea feels useful or smart, but that doesn’t always mean they actually want to change behavior right now.
Especially with career/self-improvement stuff.
This feels like a positioning problem more than a traffic problem.
The strongest idea here is not “career simulator” or “AI roleplay.” It is helping people practice high-stakes professional moments before they face them in real life. Salary negotiations, performance reviews, career decisions, manager conversations — those are moments where confidence and preparation matter.
I’d also pressure-test the name before you push more traffic. Verism is clean, but it feels a little abstract and does not immediately tell a cold visitor what kind of serious outcome they are stepping into. For a paid product asking people to trust an AI with career anxiety and professional decisions, that first impression matters.
Beryxa .com would fit this better as a serious career simulation and decision-training platform. It feels more like a product company than an experiment, and gives you room if the product expands beyond the first two simulators.
The product is already about trust, confidence, and preparation. The name should help carry that before the homepage has to explain everything.