A practical way to decide whether your software should sell itself or be sold by people, and how to run both without wasting sales time on free users.
Product-led growth (PLG) means customers find, try and buy your product largely on their own, usually through a free plan or trial, with the product doing most of the persuading. Sales-led growth means a salesperson guides buyers through demonstrations, proposals and negotiation. For software companies, PLG has become the default advice, because it promises a lower customer acquisition cost (CAC), what it costs you to win one customer, and growth without hiring a salesperson for every new deal.
It is not always the right choice. Choose product-led growth when one person can get real value from your product alone, within a session or two, at a price they can pay without asking anyone. Choose sales-led growth when deals are large, several people must agree, or the product only proves its worth after setup and integration. Many companies end up combining the two: self-service for small customers, and a sales team that steps in when product usage points to a larger deal.
Getting this wrong is expensive either way: you spend sales salaries on customers too small to justify them, or you lose large deals because nobody was there to guide the buyer. This article sets out the money behind the choice, the conditions each approach needs and what each asks of your marketing, then shows how to combine them and which mistakes make hybrid models fail.
The decision is mostly arithmetic. A sales-led deal carries the cost of people: salespeople, demonstrations, proposals and follow-up. That cost only makes sense if each customer pays enough to cover it within a reasonable time, which subscription businesses track as the CAC payback period: the number of months of gross profit from a customer it takes to earn back what you spent winning them.
Here is an illustration. Say a salesperson costs you €6,000 a month including taxes and tools, and closes four new customers a month, so each sale costs €1,500 in sales time alone. If a customer pays €40 a month, it takes more than three years of their payments just to recover that cost; if they pay €1,500 a month, you recover it in the first month.
So product-led and sales-led growth are not rival philosophies; they suit different price points. The lower your price and the simpler your product, the more of the selling the product itself has to do.
PLG reduces the cost of the first sale; sales-led selling increases the size of it.
Product-led growth works when the product can sell itself to one person at a time. Check yours against four conditions:
Figma, Slack, Notion and Calendly are the usual examples, and each meets most of these conditions: one person can start alone, and the product spreads as they invite colleagues or send links to people outside their company. If your product needs data migration, approval from several departments or a three-month implementation, a free trial will mostly produce frustrated users who never reach the point where it becomes useful.
Self-service still needs investment, just in different places: onboarding inside the product, clear pricing on your website, help content and, for European customers, a checkout that handles VAT numbers, proper invoices and several languages. Without these, users get stuck, and nobody notices, because no salesperson is there to hear about it.
Sales-led growth wins when the deal is large enough to pay for the sales effort and the purchase is too complex for one person to make alone. The typical signs are that several people must agree, the buyer's IT or security team has to approve the product, and the value only appears after configuration, integration with other systems or training.
In Europe, larger customers add steps of their own. Under the General Data Protection Regulation (GDPR), if your software processes personal data on the customer's behalf, they need a data processing agreement with you, and many will send a security questionnaire before they sign. Procurement may also ask for contracts in their language, references from their own country and invoicing on their terms, none of which a self-service checkout can handle.
Forcing a product-led approach onto this kind of product does damage. Prospects start a trial they cannot set up properly, decide the product is weak, and the market learns to expect it cheaply. As a rough rule, if your typical contract is worth several thousand euros a year and needs setup, start sales-led and add self-service later for the smallest customers.
The choice also changes what your marketing team does all day. In a product-led business, marketing's job is to bring in the right sign-ups and help them reach a first useful result: search content that solves real problems, templates and free tools, comparison pages, a clear pricing page and onboarding emails. The numbers that matter are sign-ups, activation (the share of new users who reach that first useful result) and the share who go on to pay.
In a sales-led business, marketing's job is to start the right conversations with the right companies: ads on LinkedIn aimed at named accounts, case studies with figures, events, and material that helps salespeople answer objections. Here the numbers that matter are qualified meetings, sales opportunities, win rate and deal size, and a cheap lead that never becomes a meeting is worth nothing.
Running both means running two marketing plans with different measures, so give each one an owner. A team judged on sign-ups will drift towards cheap, high-volume channels, while a team judged on meetings will neglect the self-service route. Decide which numbers each person answers for before you hire anyone or brief an agency.
Many software companies end up with a hybrid: a free plan or trial open to anyone, and a sales team focused on the users most likely to become large customers. The link between the two is the product-qualified lead (PQL), a free or trial user whose behaviour in the product shows they are getting value and could buy more.
Define PQLs by what people do, not by their job title. Useful signals include inviting several colleagues, connecting the product to another system the customer relies on, hitting a usage limit on the free plan, or several people signing up from the same company. Combine one or two of these with fit, meaning a company of the size and type you want to sell to.
As an illustration, say 1,000 people start a trial each month. Most will use the product alone and buy, or not, without help. If 60 meet your PQL rule, a salesperson contacts each of them within a working day, offering help with setup or a team plan rather than a hard pitch, and the conversation starts when the customer is already seeing value.
Review the rule every quarter. Compare the conversion of PQLs that sales contacted with similar users who were left alone; if the difference is small, the sales time is better spent elsewhere.
Hybrid models tend to fail in predictable ways. Watch for these five:
The fix for most of these is the same. Decide in writing which customers the product serves on its own and which ones sales handles, based on company size, plan or usage, and review that line every six months as your product and pricing change.