How Chili Piper narrowed 300,000 accounts to 13,000, and grew ACV by 25%

The challenge: when qualification lives in reps’ heads

Chili Piper is a demand conversion platform that turns inbound website interest into qualified pipeline, through AI agents that engage, qualify, route, schedule and re-engage.

Their customers are B2B software companies running inbound-led growth motions, and that specificity is exactly what made their qualification problem hard — because “company running an inbound motion” isn’t something any traditional database would allow you to filter by.

Here’s how Chili Piper went from a pool of 300,000 accounts and no definition of good, to a pre-qualified, highly targeted market of 13,000 accounts, and built a distribution system that gives every rep an equal shot at the best ones.

The problem with prospecting on instinct

When Tainah Subtil joined Chili Piper as an SDR in late 2021, the team had no formal ICP. Qualification lived in reps’ heads.

Some reps chased companies that had recently raised funding. Others went after accounts whose websites looked polished. A few, Tainah among them, developed a sharper intuition. Chili Piper works best for companies generating inbound leads, so the signal to look for is an active inbound motion — gated content, PLG flows or forms designed to convert.

But knowing what to look for and being able to find it efficiently were two different problems. Tainah was manually opening each account’s website, running a browser extension to check traffic volumes and separating paid from organic — about five or six clicks per account, repeated hundreds of times.

The underlying frustration was that the team’s best performers couldn’t explain why they performed well. They’d found clusters of accounts that converted, and they kept going back to similar-looking companies — but that was hard to coach or replicate, and you couldn’t stop a new rep from spending their first quarter working accounts that had no realistic chance of converting.

The scale of the problem was structural. Across the team, SDRs were touching around 5,000 accounts per month, roughly half of their entire qualified market, with no consistent way to prioritise within it. For every 23 accounts a rep worked, only one was ever going to be a good fit to buy.

The solution

Defining the real market

By analysing Chili Piper’s existing customer base, then scraping website behaviour across a broad universe of software companies, Goodfit surfaced the booking-CTA signal at scale.

As a result, Chili Piper’s addressable market went from approximately 300,000 accounts — essentially every software company in a broad set of filters — to 13,000 truly qualified accounts (later refined to 12,000).

Getting there meant surfacing 300+ account-level data points, including CRM, industry, website CTAs, geography, team size and tech stack signals.

For Tainah, the shift was disorienting at first. Losing accounts felt like losing opportunity.

Grading the market

With a defined universe of qualified accounts, the next question was how to prioritise within it. Not all 13,000 accounts deserved the same attention or the same level of resource.

• Goodfit’s grading model combined signals including:
• Website traffic volume
• Paid traffic as a proportion of total
• Sales team size
• And more

Each signal contributed to a tier, and the tiers reflect the likelihood of a rep actually booking a meeting. Tier 1 accounts are fewer, but the booking rate is significantly higher. Large, well-known companies with impressive traffic volumes are often graded lower than reps expect, because their size makes them harder to penetrate and slower to convert.

The grading model also shaped how Chili Piper thought about which accounts were worth pursuing at all. Over time, the team used their tier data to move upmarket deliberately: introducing a strict $15k minimum deal size, removing discounts, and redefining their qualified profile around companies with 100 or more employees and meaningful inbound traffic volumes.

The accounts that didn’t meet that bar weren’t chased. The ones that did got more resource, better sequencing, and reps who knew exactly why they were calling.

The grading model has been refined continuously since implementation, with Chili Piper’s revenue operations team regularly updating the signals used as they learn more about what actually predicts conversion.

Adoption climbed fast once the model was live, from roughly 50% of Chili Piper’s accounts scored by Goodfit’s algorithm to around 90% — described internally at the time as adoption having “skyrocketed.”

Step 3: Distributing accounts so every rep has an equal shot

This is where Chili Piper’s approach diverges from most teams, and where the impact of having a graded, structured market becomes most visible.

Most revenue teams score accounts. Far fewer think systematically about how those accounts flow to reps over time. The result is that experienced reps accumulate the best accounts, new reps inherit whatever’s left, and performance gaps get misread as talent gaps.

Instead, Chili Piper built a distribution system — running through their own Distro product, integrated with Goodfit grading — designed to equalise pipeline quality across the team.

Every rep carries a defined mix of tiers. A percentage of their pipeline is Tier 1, Tier 2, and so on. When an account is recycled, because a rep has been working it for six months without a booking and it’s time to move on, Distro reads the gap and routes a replacement account of the same tier. The pipeline stays balanced without anyone managing it from a spreadsheet.

New reps get a specific advantage at onboarding: the highest-graded available accounts that have been resting for at least 30 days. Fresh accounts, fresh sender reputation, and a pipeline weighted toward the accounts most likely to convert, from day one.

The system also creates a forcing function for pipeline hygiene. When pressure mounts mid-quarter and reps are tempted to chase volume over quality, the data makes it easy to audit. Filter for accounts with fewer than 10 salespeople, and remove them.

The practical consequence is that new rep performance is no longer a proxy for account quality. When every rep carries the same tier mix, performance gaps become legible — you can see whether a rep is struggling because of skill, or because they’ve exhausted a particular segment. Coaching becomes possible in a way it wasn’t before, because the variable you’re controlling for is finally fixed.

Step 4: Segmenting by signal for outbound

With accounts mapped, graded, and distributed, Chili Piper layers in a fourth dimension: campaign-level targeting based on behavioural signals.

The most powerful current example is paid advertising. Chili Piper’s value proposition — converting inbound leads more efficiently — is most urgent for companies actively spending on paid traffic. A company running paid ads has already committed to driving visitors to their website; Chili Piper’s pitch is about making sure those visitors actually convert.

Goodfit surfaces which accounts are running paid media, letting Chili Piper build targeted sequences around that specific pain rather than a generic value proposition.

A second targeting layer is competitor displacement. As Chili Piper has expanded its platform, consolidating forms, chat, and lead routing under one roof, the pitch increasingly centres on replacing fragmented point solutions.

Goodfit identifies which accounts are using specific competitors, allowing the team to build sequences tailored to the displacement conversation before a rep even makes contact.

The results

Four years after implementation, around 90% of Chili Piper’s outbound pipeline comes from accounts sourced and graded through the Goodfit model.

Roughly 20 hours a month is given back to reps who’d previously spent that time manually checking accounts, alongside an automated pool that removes 20–30 unqualified accounts per day so reps are never working a stale list.

The commercial impact of tightening the market went beyond meetings booked. With reps focused on accounts that were genuinely likely to buy, Chili Piper saw ACV increase by 25% — a direct consequence of pursuing accounts where the product’s value was unambiguous. When the fit is right, buyers negotiate less on price and expand faster.

Inbound and outbound tell different stories. Inbound converts at a much higher win rate, but at a smaller deal size, since those buyers have already self-selected. As Morgan Cliburn, who leads revenue at Chili Piper, puts it:

Closing that win-rate gap on outbound, without giving up outbound’s larger deal size, is exactly what the tiering model is built for. By making sure outbound effort goes to accounts that look like their best inbound customers, the team narrowed the win-rate gap without sacrificing deal size.