Customers I never talked to

I wrote a lot of outreach for QuickStaging. I'll get to how much of it I sent in another post. The short version: none of my paying customers came from it. They found the site on their own, signed up, and paid, almost always on the same day. The median time from sign-up to first payment is zero days.

That sounds like good news, and some of it is. It also means I never had a sales conversation with any of them.

Transactions, not customers

When I looked at what people actually bought, the pattern was obvious. An agent has a listing. They need a few empty rooms staged. They buy the smallest pack, use part of it, and leave. In my own notes I wrote it more bluntly: these aren't customers, they're transactions.

The numbers backed that up. People who bought the 30-credit pack used about 35% of it. Monthly plans sat unused and got cancelled. Credits I'd handed out outnumbered credits spent more than three to one.

So I stopped fighting it and priced for it:

• a $10 Solo plan with 20 credits, sized to what people actually used,
• a Listing Pass: 25 photos for $19, valid for 30 days from the first photo, not from the purchase. One listing, one price.

What they complained about

I never collected objections on a call, because there were no calls. What I had was the feedback box under each result. The written dislikes said things like:

stop changing the structure of the house

keep on asking to use a closet not a chest

dont understand the corrections

Furniture placement was the biggest theme, eight of eighteen written complaints. That matters more than it sounds. A staged photo that moves a wall or invents a window isn't just ugly. An agent can't put it in a listing.

What changed

Two changes are live, both about not charging people for bad output.

Flagged results are refunded. An automated quality check looks at each delivery. If it flags a result, the credit goes back, and the user keeps the photo in case it's still useful.

Failures cost nothing. A permanent error refunds at once. A capacity error keeps the credit and retries for up to about two hours, then refunds. Jobs stuck for more than 15 minutes are marked as timed out and refunded. I wrote more about these rules in a separate post.

The editing tools were built for the same complaint. A before/after slider makes it easy to judge a result. Fix area lets an agent paint the one wrong region and describe the change instead of rolling the whole photo again.

Furniture placement itself is still the open problem. It's the next thing I'm working on.

What I'd tell another founder

Inbound customers who pay on day zero are telling you they have an urgent, one-off job. Price for the job, not for the relationship you wish you had. And if you can't get people on a call, the feedback box under your output is the next best interview. Read every line of it.