William Ballance

The through-line

Ballance's fixed point is a refusal to accept the trade the rest of the sales tech market has accepted. He frames product decisions along "these two axes of what's effective and what's what's efficient," where "what's most efficient is full automation what's most effective is no Automation and intense personalization," and notes that "with high efficiency you'll have low reply rates with high Effectiveness you'll have low volume" [1]. Lavender deliberately sits in the middle: not automation, but compression, reducing "true personalization down to under 30 or 60 seconds for most users" [1]. The moral case behind it is stated from the buyer's seat: "if a sales rep won't spend 3 to five minutes to write me a personalized email and they're going to rely on some at scale component of fake personalization then why would I spend 30 minutes or an hour with them to look at their demo or listen to their sales pitch" [1]. He makes the same objection as a prospect himself, saying he hates receiving "just purely automated templated emails" and that a lot of companies "are trying to replace the sales rep and they're going like full Automation" [2].

The second through-line is that almost nothing about Lavender's growth was bought. A weekend pivot, daily LinkedIn posting, free product, free content and parties instead of booths produced enterprise adoption from the bottom up [1][2][3]. What has shifted over time is his own role: from a founder doing every single thing to a manager and recruiter, and from a two-engineer product to a company deliberately paying down years of accumulated backlog [2].

On effectiveness versus efficiency in email

The product is positioned as a Grammarly for the sales email, analysing a draft or template for what will hurt reply rates based on data, then helping fix it with AI plus workflow tools like AI personalization [1]. Crucially, the scoring is not a universal style guide. "everyone's different," he argues, so while "by and large a shorter more succinct email tends to perform better," there are situations where a longer, more in-depth email wins, and Lavender analyses historic data to work out what performed for that person or that team and recommends accordingly [1]. Different situations get scored differently. His justification for why a rep cannot do this unaided is volume: "unless they'd analyze hundreds of millions of emails they never would have this much information," and Lavender was tracking toward roughly a billion analysed emails within the year [1].

He also has a diagnosis of why professional email writing is bad, and it is not generational. People "tend to write like they were trained" in school, meaning long, formal, self-consciously educated prose padded out to hit a word limit [1]. He rejects the idea that Gen Z reps are uniquely handicapped: he was not trained on professional emails either, only on research papers [1]. What has changed is attention span, shortened across every generation by social platforms, compounded by executives whose inboxes are flooded by their team, their board and everyone selling to them, which is why cutting through usually requires concision [1]. Even if college students were trained on professional email, he adds, there would be "no no way to keep up with the amount of data and how fast the trends are changing" [1].

There is a manager-facing layer too, a coaching tool for leaders to go in and coach their team, which he describes as being to email what Gong is to calls, and which widens the user base beyond the SDRs and AEs in their twenties [1].

On the pivot and the timing that made it work

Lavender began as a startup applying communication psychology to marketing emails and social ads, and was due to launch in March 2020 [1]. Lockdown removed the pipeline and the biggest customer, a Manhattan co-working space that could not even operate, leaving a month or two of runway from angel funding [1]. The pivot was fast, an idea on a Friday and a first version out by Monday, aimed simply at people sending more one-to-one email during the pandemic, including job seekers, in the hope of making an extra ten or twenty thousand dollars to stay alive [1]. Salespeople found it instead. An early user discovered it through a Slack group, installed the beta and posted about it on LinkedIn, and it went micro-viral in LinkedIn sales communities, at which point they doubled down on LinkedIn and on salespeople [1].

He is unusually candid that the fit was partly luck, and he can name exactly which structural shift he rode. Pre-COVID roughly 70% of outbound activity was over the phone; during the pandemic that flipped, with about 70% happening through writing, in part because the phone numbers data providers sold were office numbers and nobody was in the office [1]. In the first quarter after the pandemic, outbound email volume rose about 16% while reply rates halved [1]. Layered on top was LinkedIn's own evolution from a place to host your résumé into a real social platform. He sums the combination up as "right problem at the right time with emerging social platform that no one had really cracked yet that we were able to get on really early" [1]. He also credits the original scope: because the technology was not built for sales emails alone, it turned out more effective and more flexible, and has since been fine-tuned with large volumes of sales email data [1].

On giving first

The core value he keeps returning to is stated plainly: "we give first that's like one of our core values we give first without expectation" [2]. The product is free, the content is free, and they expect nothing in return, on the theory that doing it long enough eventually converts some people [2]. The first users came from cold outbound to people laid off from Airbnb during COVID, back when Lavender was for anyone rather than for salespeople [2]. They still give the product to job seekers, most of whom get a job in something other than sales and then stop using it, while a few stay or carry it into their new team [2]. He is dismissive of the reverse-engineering this attracts, noting that people on LinkedIn dissect what they imagine Lavender's strategy was when the reality was simple and largely organic [2].

The content half of that ran on LinkedIn from lockdown onward, with him and his co-founder posting almost every day for three years [1]. His co-founder Will niched down entirely on cold email content and became "kind of synonymous with cold email," known as the email guy with 70,000 followers who follow him purely for daily cold email tips, while the Lavender brand page lagged behind the two personal profiles for a long time [1]. He is explicit about the constraint that forced this: with no money for marketing, there was no way to have built a sales team or run traditional paid marketing, so free daily posting was the only channel available [1].

On throwing parties instead of buying booths

"we've been saying internally that we don't do booths we just throw parties and that s tends to work for us pretty well" [1]. The origin was scarcity: bootstrapped with angel funding, they threw a rave in an Irish pub at SaaStr in San Mateo, planned quickly and cheaply, expecting little, and it became the biggest event of the conference and set the momentum that led to the Series A [1][2]. The pattern repeated at scale with an Alice in Wonderland themed event they called Lavender Land in a private jet hangar, over a thousand people, alcohol sponsorships, investors, customers and prospects, doubling as promotion for a streaming platform of the same name carrying in-house original content including funny shows, game shows and series [1]. It also works with almost no notice: at Dreamforce, three of them decided with 24 hours' warning to run an event, found a nightclub, and drew a crowd [1].

He is careful to say the tactic is audience-specific rather than universally clever. Lavender's users are mostly SDRs and AEs in their twenties or early thirties, often a year or two out of college, and used to that lifestyle of events, parties and clubs, so it resonates; "if I was in a different business I'm not sure this this strategy would work as well" [1]. He also stresses how thin the physical presence can be: at a conference in Dublin it was just him and one other person walking around shaking hands, no booth and no team, co-sponsoring a happy hour with partners including Chile Piper, and what surprised him was how many attendees already knew Lavender, knew him and knew about the events [1]. Their first ever booth came only recently, at a small conference in Nashville [1]. Community work has since been given an owner, with Jen Allen hired out of Challenger to build bridges with other communities, partnerships, sponsorships and organisations such as women in sales, rather than to run the LinkedIn audience [2].

On the hybrid product-led go-to-market

The go-to-market was designed around founder bandwidth. With only two people wearing every hat, a purely outbound sales motion was impossible, so they leaned on content and community, and launched something an individual SDR or AE could install and use without talking to a manager, their team, IT, or even entering a credit card, thanks to a free trial and free tier [2]. That was contrarian at the time, since most sales tech was sold top down as enterprise [2]. For a long stretch, revenue came from individual reps paying twenty and later thirty dollars a month, which pays for itself in commission if a rep doubles their reply rate [2]. The upgrade path is the multiplayer effect: once a whole team is learning from each other's data and that feeds back into recommendations, the team leader becomes a buyer, alongside the intelligence and coaching features [2]. He compares the shape to Slack [2]. The result is a genuine hybrid, with the product-led and marketing-led side generating awareness and top of funnel, and AEs running demos, pilots where needed, and closes [2]. This grassroots adoption is what carried Lavender into enterprises like Google and SAP [1].

The proof point he leads with is Twilio, which laid off 11% of staff around September 2022 while implementing Lavender and still increased demos booked by 60% compared with the previous year, results Twilio attributed back to the product [2]. His reading of that is the recession thesis for the category: teams having to do a lot with less need reps who are both more effective and more efficient [2]. Other named users include demandbase and Sendoso [2]. He is also allergic to overclaiming about the company itself: "I think we have a long way to go before I'd say we're successful but we're definitely more successful than we were when we got started" [2].

On the founder transition and technical debt

Raising money changed his job description. After the SaaStr rave they raised $2.2 million within weeks and signed a term sheet for another $11 million three weeks later, having been chronically under-capitalised with the founders doing almost everything [2]. He describes the shift bluntly, quoting his own co-founder's line that "transitioning from a builder to a leader is real," and characterises the new work as being a manager and a recruiter, putting people in the right places [2]. Running twenty interns during COVID was not the same skill, because full-time senior people are a different proposition from part-time interns [2]. The team went from six people in December to sixteen, with hiring concentrated on product and engineering, feelers out for go-to-market leadership, and a deliberately high bar for the next sales hires because inbound applications are plentiful given the brand [2].

He treats engineering capacity as both the company's Achilles heel and its proudest fact: a product used by thousands of people worldwide and by companies like Twilio, demandbase and Sendoso was built essentially by two engineers, the CTO and the first full-stack hire from 2020 [2]. The consequence is a roadmap roughly two years behind, plus bugs, over-complex flows and imperfect onboarding, none of which has stopped users getting value [2]. One user told him "all the quirks in the product it makes it more endearing" because everyone is so close to the founders, but he recognises that as the company grows, users become more removed from the founders and larger customers leave less margin for error [2]. So every sprint mixes net new work with polishing and cleanup, alongside pushing the envelope on what comes next in sales email [2]. Product investment since the raise has gone into more generative AI, including GPT-3 and ChatGPT, and expanded personalization features that let a rep research a prospect and send a personalized email without leaving the inbox [2].

On building a company from anywhere

Most of Lavender was built while he was living in Mexico, and the team was assembled inside that remote-first world, so distribution was the default rather than a policy [2]. He extended it to constant travel, all of it tied back to the business: Toronto, where meeting a filmmaker led to the Lavender Joe videos about a cringy salesperson doing cringy things to real people; SaaStr in San Mateo; Web Summit in Portugal on a speaker's guest pass sourced through a LinkedIn contact; an Outreach partner conference in London; an angel investor in Barcelona; Slush in Finland, reached via a free ticket from that investor after a chance hotel-lobby conversation; and Poland and Prague as guests of Woodpecker, an email automation company he had used personally for years [2]. He signed the Series A term sheet while in Portugal, which made the rest of the trip better and also began a month of diligence and lawyers [2].

On the practicalities, only Finland's eight-hour gap really hurt, with team meetings at 11pm and some shuffling of times by a flexible team [2]. He argues the time difference had a genuine upside: a whole morning to himself before anyone is awake to DM, email or text, then the team coming online around midday, making for a longer but more balanced and more fulfilling day than the US pattern of waking up and being immediately on until the day is over [2].

Takeaways

• Build against two axes, effectiveness and efficiency: full automation is the most efficient and least effective, and the winning position is compressing genuine personalization to under 30 to 60 seconds rather than faking it at scale [1].
• The buyer's logic against automated outreach: if a rep won't spend three to five minutes writing to you, there is no reason to give them thirty minutes of your time [1].
• Shorter emails usually win, but not always, so score every draft against that individual's and that team's own historic reply data rather than a universal rule [1].
• COVID flipped outbound from roughly 70% phone to roughly 70% writing, while post-pandemic email volume rose about 16% and reply rates halved, creating the exact problem Lavender solved [1].
• Give first without expectation: free product, free daily content, and product given to job seekers who mostly never become customers [2].
• Don't buy booths, throw parties, but only if your audience is young reps for whom that lifestyle resonates; a bootstrapped rave at SaaStr preceded a $2.2 million raise and an $11 million term sheet [1][2].
• A hybrid motion works when an individual rep can adopt the tool for $20 to $30 a month with no manager, no IT and no credit card, and team-level data makes it a manager purchase later [2].
• A tiny engineering team can build a beloved product and still leave you two years behind your own roadmap; expect to spend post-funding sprints on cleanup as enterprise customers reduce the margin for error [2].