Launched today: built an alternative to spray-and-pray outbound. Here's the stack, the burn, and the day-1 numbers.
Hey IH 👋
Solo-ish founder here (me + one engineer). Just pushed Backchannels live this morning after 10 months of building.
## The problem I'm trying to fix
Cold outbound is broken at a fundamental level and nobody wants to say it out loud.
The standard playbook:
1. Build a list of "anyone with VP/Director/Head of in their title at companies matching your ICP"
2. Run 100K emails through a sequencer
3. Get a .4% reply rate
4. Call it a quarter
We've decided this is fine because it occasionally works at scale. It's not fine. You're paying SDRs $80K/yr to spray emails at people who have no authority to buy what you're selling.
My last team did this for a year. We sent ~85,000 cold emails. Booked ~340 meetings. Of those, ~40 were with actual decision-makers. The rest were curious bystanders, junior researchers, and one person who thought we sold something we didn't. Closed ~12 deals.
The math: ~$2,100 per closed deal in SDR cost alone - not counting the $24K list contract, the Outreach license, or the consultant we hired to make this work.
I kept thinking: if we'd just emailed the right 4 people at each account instead of any 12, we would have closed more with 1/10th the labor.
## What I built
Backchannels identifies the actual buyer inside each target account, instead of returning "anyone with a matching title."
- 225,000 verified SaaS decision-makers
- Each contact carries decision-authority signals (role, buying behavior, budget indicators, hiring patterns)
- Filter by company, role pattern, tech stack, intent
- Preview free, pay $0.08/contact only on unlock
- Direct sync to Salesforce + HubSpot
- No subscription, no contract
## The build
Stack:
- Typescript, Node.js, TypeORM, Next.js, PostgreSQL
- Heroku
- Public sources + partner data + light scraping for the verification layer (won't pretend it's all proprietary)
- Stripe for billing
- Salesforce + HubSpot APIs for sync (took longer than everything else combined)
Timeline: 10 months. ~5 months on the data pipeline + buyer-identification model. ~6 weeks on the actual app shell. ~2 months on integrations and polish.
Costs to date:
- Hosting + data: ~$2,200/mo running cost
- Contract work: ~$6K total
- Legal (entity, contracts): ~$3K
- Total out of pocket so far: ~$31K
## The pricing experiment
Most contact databases charge $14-26K/yr per seat. I'm charging $0.08 per contact unlock, no commitment.
This works because there's no enterprise sales team eating 30-50% of revenue, no CS overhead sized for $25K accounts, and CAC is structured around an $80 entry purchase. The tradeoff: probably uninvestable on the standard SaaS playbook. I'm committed to bootstrapping.
## Day 1 numbers
7 paying customers from a 3-week closed beta. Combined revenue: $680. Today is the open launch. No huge audience to drop this on.
## What I need from this community
For IH folks who've run outbound for a SaaS - what's your real reply-to-close ratio? I'd love to gut-check whether my math is representative or whether I'm anchored to my own bad experience.
And for anyone building usage-based pricing in a subscription-dominated category:
- How did you handle procurement at larger customers wanting POs?
- What's your churn metric when there's nothing to actually churn from?
Site: backchannels.co
Will post a 30-day update with real numbers, conversion rates, and whatever broke.
The useful angle here is not just virtual try-on.
It is reducing the hesitation moment on fashion product pages: “Will this actually look good on me?”
That is the pain store owners understand.
I’d be careful not to position ETRYON as only a cool AI feature. The sharper test is whether merchants see it as a conversion lift tool for products where fit, style, and visual confidence block purchases.
That is probably the wedge worth testing first.
I think, as small business owners most of us have been on both sides of this. Trying to market a product or service, and managing the never-ending inflow of cold outreach to our business for products or services offered to us. For me, cold emails have never been worth the cost, and mostly ineffective.
What has worked is a combination of more traditional advertising and actual conversations and networking. And this is the only method that has worked for products or services offered to me as well that did not end up in my spam email box.
I think, however, this gets more challenging for Solo Founders and SaaS products, as some of the "closer to home" channels (local chamber of commerce, local libraries, senior centers, etc.) that worked for my small IT business don't scale.
I am new to this space so I was glad to find Indie Hackers — it's been a wealth of real-time knowledge. I am eagerly reading the other replies here and am glad to see someone posting what I was curious about myself.
You're right that networking and traditional advertising have a track record cold outbound can't touch - they start from trust or genuine fit, not interruption. That's the part that gets glossed over in the "cold email is just a numbers game" discourse.
What I'm trying to do with Backchannels isn't to replace what worked for you. It's to narrow the gap. If outbound goes to the ~3 humans per company who actually have buying authority (not the 17 with matching titles), the outreach starts looking less like spam and more like a targeted intro from someone who actually researched whether you'd care. Still cold, but no longer random.
The relationship still has to be built. The product doesn't fix that. It just makes sure you're building it with the right person. Indie Hackers is a great community for this kind of conversation - glad you found it.
60 percent+ open rates on cold email by ditching templates and writing one specific data point about each recipient. What were your day-1 conversion numbers, and how are you thinking about reply quality vs reply quantity?
Always Quality > Quantity. If anyone says otherwise they know nothing.
Congrats that looks good but open rates on emails can also be flawed stat as many email providers at the corporate level auto-open emails to scan for malicious emails so I would also check if your SEP includes those in their open rate metric. AKA was it human open or a system.
Do you only focus on replies or do you also measure click rates on links?
Feels like people want real conversations again instead of endless cold outreach.
Very true and the Cold Outreach approach doesn't lead to that anymore. The system is all geared towards massive volume with a small trickle of success.
Conversations don't happen when those emails are in your spam folder as well
The 85K → 340 meetings → 40 actual decision-makers math is exactly the part of the cold outbound conversation nobody publishes. Most teams I see still optimize for top-of-funnel volume because that's what their tooling shows them.
The thing that should kill spray-and-pray faster than tooling: B2B buying committees are 6 to 10 people now, so "find the VP" is the wrong frame entirely. You need depth across the committee, not breadth across companies. Curious how Backchannels handles the multi-threading piece. The SDRs at every Microsoft partner I worked with for the last decade still treat outbound as one champion per account, and that's where the math actually breaks.
You're identifying something more important than what I led with. "Find the VP" is genuinely the wrong frame now - research puts the average B2B buying committee at 6-10 humans and climbing every year. Single-threading those accounts is where the math actually collapses on the enterprise scale.
Honest answer on multi-threading:
Today: we identify the primary decision-maker per account with role and authority context, and the same query surfaces other likely stakeholders (titles, departments, reporting structure). You can see and unlock the adjacent set in one motion.
What we don't do yet - and what your comment is rightly pointing at - is explicit committee mapping that says "these 6 people are the buying group, here are their roles in the decision (economic buyer, technical evaluator, champion, blocker)." That's a sophisticated problem still takes a proficient seller to be able to intertwine.
We're trying to compress that whole motion to one query (we give you buyer of your competitor for example) so you can have singular focus and then branch out into multi-threading as you engage the buying committee.
The Microsoft partner observation is a great gut-check - that ecosystem has some of the most complex buying motions in enterprise SaaS. Would genuinely love your take on what "good" multi-threading looks like there. DM me if you're open to a 15-min chat - always trying to calibrate the model against the harder cases and your genuine feedback on what I've built.
Looks good . congrats . i will give it a try and see
Thank you please let me know any feedback you have
sure
Respect the bootstrap. How are you differentiating the buyer-ID from ZoomInfo/Apollo/Hunter, and how do you measure when you're wrong?
Thanks. Two halves to this answer:
Differentiation: ZoomInfo/Apollo/Hunter are optimized for breadth - give me everyone with "VP" in their title at companies matching my filters. They're great at that. We're optimized for *depth on the buying decision* within an account, which 2-4 humans actually purchased and been a buyer with our proprietary dataset.
* Basically they are guessing who the best contact is on the buyer/ using Intent data but the accuracy of intent is still questionable. Intent doesn't equal actual buyers (some buyer behavior) although it can be directionally correct.
We haven't been wrong yet according to our customers but we do allow for customer feedback within the product so if it is wrong then it will be flagged and human researched + corrected.
What has your experience been with these platforms around souring buyers?
This sounds amazing. I have been in outbound Sales for a few years and knowing exactly who is part of the decision committee makes my job a whole lot easier and save me weeks. I just bought $80/month worth of contacts. Think you need to update ur revenue numbers now 🙂
Thank you I appreciate the support and hopefully this will make your life a little bit easier
Looking to add a design engineer to the team?
Potentially in the near future, what would you recommend for initial improvements?
I've thought of these problem in a different market but I believe besides exceptional targeting powered by scraping and the messaging could also be narrowed down future with autofill that makes the sequence messages personalized beyond just tokens.
And data doesn't only have to come from the built in contact list, since in my case I was followed on simplifying sequencing and task delegation with teams and agents, I tend to believe that enterprise users could draw in data from any of their entry points into a single that database that could be analyzed and marketed to.
I tend to think of this whole thing in terms data source, operations and feedback.
All software do the same, collect data from one source and pass to another. With that in mind I designed Monolith to integrate with the users apps, so they set up their sequences, pull that from those platforms, and run operations [sequences], with agent and teams to that can work copy and data analysis independently.
This what am actually building rn. Interested in seeing how you might think of them.
Appreciate the framing — data source / ops / feedback is a clean way to think about it.
I'd argue Backchannels and Monolith are solving different layers of the same stack. We're upstream of where you're focused. Our thesis is that contact identification is where most cold outbound fails - before messaging or sequencing even gets a chance to work. You can have perfect autofill, AI-written copy, and clever cadences, but if the sequence is going to someone who can't actually buy, you're just running more sophisticated spray-and-pray.
The order of operations is target → automate. Most tools (incumbents included) try to automate first and target second. That's a big part of why outbound reply rates have collapsed across the industry.
There's probably a world where something like Backchannels feeds the identification layer and Monolith runs workflows on top. I'm interested in what you're building though.
thats some great insight
Understandable, I just gave you a follow on X. you follow back and send a DM if you'd like to engage one on one.
Great just did