HackerNews中文版

早在 2015 年，我们推出了第一个 SaaS 产品，当时没有风险投资，经验也很少。大多数定价选择都来自于直觉，或者是一些我们不该听取的建议。现在回想起来，今天看来显而易见的事情，当时却并非如此。我们是付出惨痛代价才学到的。分级定价。从纸面上看，这似乎很公平。一个有 15 个用户的团队支付 39 美元，但再增加一个用户，价格就跳到了 119 美元。在同一级别内，客户对每用户成本下降感到满意。但跨越级别会让他们非常愤怒。人们流失、要求折扣或发送愤怒的邮件。这种模式没有奖励增长，反而惩罚了增长。按用户计费。最终，收入随着客户的增长而增长。但这吸引了许多单人用户和小型团队。由于 Everhour 与项目工具同步，即使是一个自由职业者也可以用数百个项目和任务淹没系统。他们几乎不付费，却造成了巨大的负载，而且是最吵闹的群体：愤怒的推文、退款请求、无休止的工单。而大型团队则更冷静、更忠诚。为了解决这个问题，我们设置了 5 个用户的最低限额。你可以使用更少的用户，但仍然需要支付 5 个用户的费用。这过滤掉了自由职业者，提高了我们的平均客单价，并让我们与合适的客户保持一致。有些人抱怨这太贵了，但那些重视产品的客户留了下来。后来，我们增加了一个免费计划，最多可供 5 个用户使用，不提供集成。这为小型团队提供了一种继续使用的方式，并让我们被列入“顶级免费工具”博客。它更多地起到了营销作用，而不是盈利作用。我们 Trello 插件的统一费用定价。每月 10 美元，无限用户和项目。在超过 30,000 个活跃用户中，只有大约 500 人付费。其他人在我们要求付费的那一刻就离开了。该产品过于简单，无法进行追加销售。祖父条款。如果有人注册了旧的计划，他们可以永远保留它。没有强制升级，没有意外涨价。许多人至今仍在付费。我们试图通过提供额外功能和忠诚度折扣来引导他们使用新计划，但只有一小部分人切换了。如果他们的计划运行良好，那么让他们支付更多费用的障碍就太大了。折扣。大客户的批量折扣效果很好，有助于达成交易。“在 5 天内注册可享受 8 折优惠”之类的紧急促销活动带来了一点转化提升，并提高了客户留存率。这些客户留存的时间更长，因为他们不想失去他们的优惠。但这对现有用户来说感觉不公平。有些人甚至试图取消并重新订阅。我们最终停止了这种做法。计费机制的教训。首先，我们按用户计费，并将按比例分摊的费用推迟到下一个周期。理论上很合理，但实际上却很混乱。客户在续订前取消了订阅，我们损失了钱，尤其是在年度订阅上。其他人无法识别银行对账单上的总额。我们尝试在添加用户时立即收取按比例分摊的费用。这解决了收入问题，但产生了太多微小的付款和发票。客户讨厌在更换用户时多次收费。最终改为按席位计费。提前购买席位，灵活使用，发票保持清晰。仍然不完美。有些人忘记删除空闲席位，后来抱怨。但总的来说，它干净多了。今天，这基本上是 SaaS 的标准。经验教训。- 分级定价看起来很公平，直到你跨越级别。然后它会让客户生气。 - 单人用户使我们的系统超载，但收入却很少。 - “微型”计划增加了支持成本，但没有带来转化。 - 统一费用定价吸引了 3 万多用户，但几乎没有人付费。 - 没有最低席位意味着客单价很低，而且目标客户不对。 - 多个计划增加了复杂性。大多数人最终还是选择了同一个计划。 - 向祖父条款用户追加销售失败了。如果旧计划运行良好，他们不会支付更多费用。 - 折扣促销活动有助于短期，但感觉不公平，并且被滥用。 - 按用户比例分摊费用使发票变得混乱，并让我们损失了钱。 - 免费计划提高了营销可见度，而不是升级。

查看原文

Back in 2015 we launched our first SaaS with no VC and little experience. Most pricing choices came from gut feeling or advice we shouldn’t have taken. Looking back, what feels obvious today wasn’t obvious back then. We learned the hard way.Tiered pricing.On paper it looked fair. A team with 15 users paid $39 but adding one more jumped it to $119. Inside a tier customers were happy as cost per user dropped. But crossing a tier made them furious. People churned, demanded discounts or sent angry emails. Instead of rewarding growth the model punished it.Per-user billing.Finally revenue scaled with customers. But it attracted many solo users and tiny teams. Because Everhour syncs with project tools, even one freelancer could flood the system with hundreds of projects and tasks. They paid almost nothing but created huge load and were the loudest group: angry tweets, refund requests, endless tickets. Larger teams were calmer and more loyal.To fix this we set a 5-seat minimum.You could use fewer but still pay for 5. It filtered out freelancers, raised our average check and aligned us with the right customers. Some complained it was too expensive but those who valued the product stayed. Later we added a free plan for up to 5 users without integrations. This gave small teams a way to stick around and got us listed in “Top Free Tools” blogs. It worked more as marketing than monetization.Flat-fee pricing in our Trello add-on.$10/month, unlimited users and projects. Out of 30,000+ active users only ~500 paid. Everyone else left the moment we asked for money. The product was too simple for upsells.Grandfathering.If someone signed up on old plan they could keep it forever. No forced upgrades, no surprise hikes. Many are still paying today. We tried nudging them to newer plans with extra features and loyalty discounts, but only a small percentage switched. If their plan worked fine, the barrier to paying more was too strong.Discounts.Volume disc for large clients worked well and helped close deals. Urgency promos like “Sign up in 5 days and get 20% off” gave a small conv boost and improved retention. Those customers stayed longer because they didn’t want to lose their deal. But it felt unfair to existing users. Some even tried canceling and resubscribing. We eventually stopped.Billing mechanics lesson.First we billed per user and deferred prorates to next cycle. Logical in theory, messy in practice. Customers canceled before renewal and we lost money, especially on annuals. Others didn't recognize totals on bank statements. We tried charging prorates immediately when someone was added. That fixed revenue issues but created too many micro payments and invoices. Customers hated multiple charges when swapping users.Eventually moved to per-seat billing.Buy seats in advance, use them flexibly and the invoice stays clear. Still not perfect. Some forget to remove empty seats and complain later. But overall it’s far cleaner. Today it’s basically the SaaS standard.Lessons.- Tiered pricing looks fair until you cross tiers. Then it makes customers angry. - Solo users overloaded our system while bringing little revenue. - “Micro” plans added support costs without conversions. - Flat-fee pricing attracted 30k+ users but barely any paid. - No minimum seats meant tiny checks and the wrong audience. - Multiple plans added complexity. Most picked the same one anyway. - Upselling grandfathered users failed. If old plans worked, they wouldn’t pay more. - Discount promos helped short term but felt unfair and were abused. - Per-user prorates turned invoicing into a mess and cost us money. - Free plans drove marketing visibility, not upgrades.This is near the 4,000 char limit here. More details and screenshots: https://medium.com/@citizenblr/lessons-from-10-years-of-saas-pricing-experiments-4ed45f552171

十年，十个定价错误：我们无VC融资打造SaaS的经验教训