The Cold Call Is Coming Back — and Cross-Border Sellers Should Be Paying Attention
Here’s a number that should bother every cross-border operator reading this: 80–85% of cold calls don’t get picked up. That’s not a stat about failure — it’s a stat about where the leverage sits. If four out of five dials die at the pickup stage, then the five seconds after someone says “hello” are worth more than any subject line, any ad creative, or any TikTok hook you’ll test this quarter. That’s why I paid attention to throxy, a Y Combinator alum that just launched local-language calling aimed at B2B outbound. On the surface it’s a sales tool for SaaS companies. Underneath, it’s a case study in something cross-border sellers have been slow to internalize: language is a conversion lever, not a localization checkbox.
What throxy Actually Solves (and What It Doesn’t)
Let me be precise about what this product is, because the Product Hunt framing (“local-language calling”) undersells the interesting part. throxy is a done-for-you outbound calling service. When a lead comes up, the system checks where that person is and routes the call to an SDR who speaks their language. The SDR opens in that language. As co-founder Pablo Jimenez De Parga puts it in the launch post, it matters most when the person picking up runs a factory floor in Lyon and has never taken a cold call in English.
The numbers they’re citing: 149 companies served since finishing YC last spring, more than 4,500 meetings booked, and over 3.8 million cold calls made, with a stated next milestone of 10 million calls (a target the team says is roughly a month out). They’re also running their own dialer, called Maestro, which routes a picked-up call to the best caller for that specific product and that specific language, with live in-call coaching layered on top.
What it does not do, and this is important for anyone reading this as a product review rather than a strategy essay: throxy is not self-serve software. There’s no dashboard you log into and start dialing. It’s a service engagement, and the way you test it is a one-month pilot — a point that came up directly in the comments when a user asked for a “call me back” button to judge call quality. The team’s answer was honest: they’re building custom databases per client, sourced from niche offline data, and the pilot is the evaluation mechanism.
Why this is an outbound-infrastructure story, not a SaaS story
The interesting thing here isn’t the language routing. It’s the fact that a company built a proprietary dialer, a routing layer, and a per-client data build before it built a self-serve front end. That’s the shape of a business that believes the moat is operational, not interface. For cross-border sellers, that’s a tell. The tools that actually move your numbers in 2025 are the ones that take a messy human workflow — dialing, pickup, qualification, follow-up — and industrialize the routing inside it.
How It Stacks Up Against What You’re Already Using
If you’re running DTC or Amazon FBA, your outbound stack probably looks like Klaviyo for email flows, maybe Attentive for SMS, and increasingly some AI SDR tool bolted onto HubSpot or Salesforce for B2B wholesale outreach. The AI SDR category — think 11x, Artisan, and the long tail of lookalikes — has spent two years selling one promise: more volume of digital comms. More emails, more LinkedIn touches, more sequences.
throxy is making the opposite bet, and the launch thread makes that explicit. When a commenter noted that AI SDRs are just adding to the noise in already-saturated email and LinkedIn channels, the throxy team agreed — they said they noticed those channels becoming much more saturated about a year ago, and that cold calling is the part of the job SDRs hate most, which is exactly why a specialist doing it well has an edge.
That’s a real competitive position, and it’s the same argument I’d make to any cross-border seller who’s been pouring budget into Meta and TikTok acquisition while ignoring wholesale and B2B channels entirely. If your CAC on paid social has crept up 30% year over year and you haven’t tested a single outbound motion to independent retailers, distributors, or brand partnership targets in your category, you’re leaving a channel unexamined.
Why Amazon sellers should care more than Shopify ones
Here’s my contrarian take: for pure Shopify DTC operators selling consumer goods to consumers, local-language cold calling is mostly irrelevant. Your buyer isn’t a plant manager in Lyon. But if you’re an Amazon FBA brand owner who has graduated from single-SKU arbitrage into brand building, you almost certainly have a B2B problem you haven’t named: wholesale accounts, distributor relationships, and marketplace expansion into EU and LATAM where your category manager or retail buyer is a human being who answers a phone in their native language. That’s the exact use case throxy is built for, and it’s the one most Amazon-native operators have zero infrastructure for.
The same logic applies if you’re scaling on TikTok Shop into new geographies. Your creator and affiliate outreach is a form of cold outbound, and the language of the first ten seconds matters just as much there as it does on a dialer.
What Cross-Border Sellers Can Borrow From This
Forget buying throxy for a second. The transferable lessons are more valuable than the product itself.
First: language is a pickup-rate variable, not a polish variable. The launch thread has a comment that stuck with me — a plant manager in Lyon picking up and hearing French from the first second is so obvious once you say it, but almost nobody does it. That’s the whole insight. A cold call in someone’s second language is over before the pitch starts. Cross-border sellers should be auditing every outbound touchpoint — email subject lines, WhatsApp templates, creator DMs, distributor outreach — for whether the first line is in the recipient’s language, not whether the landing page eventually localizes. If your first impression is in English and your localization kicks in at checkout, you’ve already lost the pickup.
Second: the “pay per meeting” model is worth studying even if you never buy it. A commenter asked the sharp question — who decides if a meeting counts as qualified when client and vendor disagree? The answer: it’s always up to the client. Before the meeting, throxy sends the full context of the conversation, including qualification questions asked and company context, and the client says yes or no. Only once the prospect actually attends is it billable. That’s a clean alignment mechanism, and it’s the same logic you should apply to any agency, affiliate, or influencer relationship you’re running. Pay for attended outcomes, not impressions.
Third: the “signals before they’re public” argument. When asked about intent-signal tracking, the throxy team said they focus less on tracking intent and more on opening conversations — because at scale, cold calling lets you hear when a company’s contract with a specific vendor is up, or before they make a specific hire, which is a whole additional layer of value. That’s a reminder that in cross-border wholesale and distribution, the best market intelligence still comes from conversations, not from dashboards.
Where the math breaks
Two honest caveats. The 80–85% no-pickup rate cuts both ways — it’s throxy’s argument for why pre-call research is a waste of time, but it also means the economics only work if your per-meeting price is low enough to absorb that waste. And the “we work best with B2B software companies” framing in the launch post is a real constraint. If you’re a consumer goods brand, you’re not the target customer, and the playbook won’t map one-to-one. Borrow the principles; don’t assume the product fits.
Where My Judgment Says This Falls Short
I’ll be blunt about the gaps, because the launch thread itself surfaces them.
There’s no self-serve test. The commenter who asked for a “call me back” button to judge call quality got a polite “great idea” and a pointer to the one-month pilot. That’s a legitimate answer for a service business, but it means you cannot evaluate throxy the way you’d evaluate a Helium 10 trial or a Klaviyo sandbox. You’re committing to a month before you hear a single dial. For a cross-border seller used to free trials and freemium, that’s friction.
The qualification definition is client-dependent, which is both the feature and the risk. “It’s always up to the client” sounds clean, but in practice it means the vendor has limited incentive to push back on a meeting you might call marginal. The context package they send before the meeting is the safeguard — read it carefully, and set your qualification bar in writing before the pilot starts.
The product is explicitly B2B-software-shaped. The launch post says they work best with B2B software companies that have a working product, a clear buyer, and a market that’s hard to reach with email sequences alone. Cross-border physical-goods sellers are adjacent, not core. If you’re an FBA brand trying to reach EU distributors, you may be a fit — but you’ll be a use case the team is still learning, not a use case they’ve optimized for across 149 clients.
And the “cold calling is back” thesis is a bet, not a law. The commenter who raised it framed it as something they’d been hearing from people they respect — not settled fact. I’d treat it the same way. Cold calling is back for certain categories with certain buyer profiles. It is not back for everyone, and anyone selling you a universal “outbound is the new inbound” narrative is selling you something.
What I’d watch / test next
Three concrete things you can do this week, regardless of whether you ever talk to throxy.
One: audit your first-line language. Pull your last 50 outbound emails, WhatsApp templates, and creator DMs to non-English markets. Count how many open in the recipient’s language versus English. If the number is low, that’s your cheapest conversion test of the quarter — rewrite the first line only, keep everything else constant, and measure reply rate.
Two: run a one-market outbound pilot on a channel you’ve ignored. Pick one EU or LATAM market where you have wholesale or distributor ambitions, build a list of 100 targets, and test a localized first-touch motion — phone, email, or WhatsApp, whichever your buyer actually answers. Measure attended conversations, not sends.
Three: pressure-test any “pay per outcome” vendor on the qualification definition. Before you sign anything with a pay-per-meeting or pay-per-lead model, write down your qualification bar and require the vendor to send the full context package before you accept. The throxy model is a good template — copy the mechanism even if you don’t copy the vendor.
The broader lesson: cross-border selling has spent a decade optimizing the digital surface of the funnel — landing pages, ad creative, checkout localization — while leaving the human surface (the first ten seconds of a real conversation) almost entirely unoptimized. throxy is a reminder that the human surface is where the next round of margin is hiding. Watch the category. Test the principle. Don’t buy the pitch until you’ve heard the pickup.
