See all the dumb things we did with our b2b SaaS pricing and the “obvious” pricing model that worked, doubled revenue and made us profitable.
- Key lessons learned
- The model that worked
- All the models that failed and why
Key lessons that got us to a pricing model that worked
- If 10 employees in a client's company are using it 40 hours each week, it's worth more than $50 per month!
- Knowing your user's job titles and how much they use it helps figure out what value they get from it and why.
- Build and look at your customer health scores. It showed us who our heaviest users were and who hardly used it. We stopped bullshitting ourselves about our real market.
- Be very careful with or totally avoid doing unlimited use plans.
- Don't add any price to enterprise plans because each enterprise deal varies too much and may take far more work than expected.
Some context: We're bootstrapped in a new market
Upscope co-browsing is the new interactive screen sharing for supporting users like never before.
It's now a profitable and bootstrapped SaaS company.
The market for co-browsing is still relatively young, and providers vary in the way they provide co-browsing and the tools that come with it.
I don't know how being bootstrapped would make a difference. If we had funding maybe we'd have different priorities on the speed at which we generate revenue and fund new feature development. Maybe not.
In short, the options on pricing were wide open when we started and we could only partly go off other companies in the same market.
Here's our current pricing that works
Things to note:
- It's priced per agent with a minimum of 3 agents.
- There's a starter plan, a business plan and no pricing mentioned on enterprise.
- Features for the business plan are made for larger companies who need them.
We tried every other dumb pricing model and this was the winner.
It was the winner because we are charging 1/10th of the value they get from it so we can be confident in our pricing.
It works because it charges the right companies for the value they gain from the service.
We're not charging those who partly need it, the pricing is aimed at those who really get good value from it. We happily tell others that they might not need it.
We figured this out by doing multiple experiments and analysing individual users to see what value they got from the product.
After the experiment and analysis the pricing became a lot simpler and more importantly, we had confidence in it.
Before we launched
March 2016 Beta version running for free for a full year.
Why a full year you ask?
It's co-browsing which is hard tech to get right.
Co-browsing is no-download instant screen sharing. Remember when you were too lazy to set up a screen share with that customer and just typed instructions? This is instant because you pass html/css rather than stream a video so there are no downloads.
It's cool but you have to build it to work with all the different frameworks that websites are built with these days.
March 6th 2017 started commercial version building.
Launch planned for May so it was time to make up pricing.
Dumb pricing model 1 that failed
We decided on small, medium, large, extra large plans.
They were divided according to the number of concurrent screen shares per company.
This meant that on the medium plan, if more than 3 people were trying to screen share at the same time, it would stop the 4th and say you have to upgrade.
What was wrong with this model?
In our previous startup we had 4 pricing plans and that worked.
We repeated this here even though it's an entirely different product.
The main reason though is that very few companies have enough staff to be doing e.g. 20 concurrent screen shares. They might use it heavily but not all at the same time.
However, we did not figure this out just yet.
See what happened 10 days later..
Even dumber pricing model 2
May 2017 - 10 days after the previous pricing change
We switched to our unlimited agents plan.
The only limits were on active concurrent users and total users per day.
Why did it fail?
Unlimited anything means we did not value our time or our product correctly.
It meant that Apple could sign up, use it for 10,000 hours per month and pay peanuts.
But they wouldn't sign up anyway.
They would not have signed up because they don't trust any company offering a service so cheap.
A manager looking at that pricing would probably think we're either too dumb to survive or too cheap to trust.
We've tested offering unlimited seats or credits across 2 different SaaS companies and sales actually dropped.
They not only trusted the service less but they asked more questions and the few companies that did willingly sign up had 30 agents using it and were costing us money while getting 1000s of dollars of value from it.
First signs of hope with pricing model 3
We introduced monthly agent limits. 1, 2, 5, 10. So, no more concurrent agents or unlimited agents nonsense.
This is closer to our current plans but there was no difference between the plans apart from the number of agents that could screen share in that month.
More hope with a small change in pricing model 4
June 14th 2017
We added the activity log and only gave it to plans that were $99 or more.
Now we had at least some features based differences between the plans.
Even more hope, then back to being dumb in model 5
Moved single sign on and user roles to higher plans and called it "startup" and "business".
Startup plan had 3 agents.
But the business plan went back to being unlimited (max 50 concurrent) agents.
In other words, we went back to being dumb.
Why did we do this?
We wanted to separate small and big and encourage them to go to the big plan.
Totally dumb pointless change to pricing in model 6
March 2018 - we doubled pricing.
Starter plan of 3 agents $99. $199 for the unlimited plan.
Why did we do this? Why did it fail?
We doubled pricing because we figured customers were getting more than 10X value and we worried we were not charging enough.
Why double pricing? We read it somewhere so we did it.
It failed because when there's no reason behind something then it's hard to have confidence in it, there's no trust in the results even if they do come about. Even if they did rise, the next downward shift would make us anxious.
We saw ups and downs in sales and we had no why they happened.
It was around now that we really began to analyse our users in more depth.
Came back to our senses in pricing model 7
April 2018 - we started to come to our senses possibly after our initial push into proper evaluation of usage and of course the lessons learned from before.
$48 up to 5 agents. $99 for 15 agents.
There were still 2 problems with this.
First, it's too cheap. If a company has 15 staff using it 150 hours in a month for converting and retaining customers, it's worth more than $99 per month.
Also, if you needed 6 agents you had to go to the higher plan and customers felt like they were paying too much for 6 users. The product might still be cheap for them but that's how much expectation and those perceived cost signals matters.
We finally knew enough to get it right in pricing model 8
Our current plans, give or take a few tweaks, are the image posted further above but if you're feeling lazy, here they are again.
We have a starter and a business plan.
Both have 3 agents with the business plan costing more as it has more features.
If more than 3 agents use it, it automatically adjusts and charges for it.
The extra business plan features are aimed at larger companies that need things like 'role based access control' and a full audit log.
Also note that the pricing is written per agent so people know what it costs per person. It cuts down on the confusion.
Why did this double revenue? Why is still growing strongly?
From the point that we introduced this pricing, our revenue doubled. It doubled because when people are getting 10X value from it, they don't need a super cheap deal to keep them onboard. They want great support, great uptime but as long as you're not charging the earth, it's still a bargain. Also, our largest clients gave it out to more team members and per seat means we expand as they expand. Our previous pricing looks like madness compared.
A further note on per seat pricing. Early on, customer churn was more important but over time and due to per seat pricing, revenue churn becomes of greater importance. To see why, have a look at our growing understanding of SaaS metrics.
We'll explain the enterprise pricing further below.
What pricing philosophy eventually worked for us?
If they're getting 10X return on investment, they'll pay for it and stay with us so price based on value.
We thought "What would we pay for Upscope if we used it that much?" We were already paying incredible amounts for tools like Intercom and not complaining.
Bad fit customers will cancel so build for those who do get a 10X return. Proudly reject those who don't need it or we'll both get hurt in the long run.
What specific actions helped us work out the pricing?
I guess you do have to do experiment to figure things out because early on you don't know enough to figure out the right pricing. As we discovered in our pricing analysis of 25 other companies, it helps to experiment faster.
Analysis of job titles
We analysed who signed up first and what their job title was. We used SEMRush to figure out how big they were and what their motives and concerns might be.
We analysed the people who used it the most. The people signing up first were rarely the heaviest users, they were often managers or higher, those who read blog posts like this, who first decided to try the product. The team members added later were account managers, customer success, support and others. They used it a lot.
We figured out exactly who used it and why by also directly asking them.
It helped us understand their time and what that might be worth and the results they got from Upscope in that time.
Creating health scores
We created a health score dashboard.
It's a list of all our paying customers, how much they use it and their overall "health" using a combination of factors to decide if they are getting value from the product.
How did we figure out our enterprise pricing?
We wrote up our enterprise pricing lessons including a breakdown of 3 deals here.
Below we'll cover some key lessons learned.
We initially wrote that enterprise pricing starts from $800 but that turned out to have little to do with reality.
Each large customer had entirely different needs and the amount of work each one took varied enormously.
They're not experimental new monthly subscribers, they're looking for a deep, long running relationship based on trust and proof of value.
The way we have figured out pricing for enterprise clients so far is to talk to them, figure out their ROI based on their specific needs and price accordingly.
Also, there are many types of enterprise clients.
There are those that just want our standard product but for a lot of support agents. Their prime concern appears to be trust and a way to prove ROI within their department and beyond.
There are others who wish to integrate Upscope co-browsing as part of their own feature set and that requires putting yourself in their business shoes and working with them to help them understand the ROI as well as pricing in stages so they're paying more as they get more use across their expanding user base. The pricing might have to be per seat or usage based. It depends on what stage they're at and what sort of company they are.
We were lucky with enterprise pricing because we were lucky with our attitude. We didn't know any better so we focused on getting them a result and understanding how they themselves might measure success. We gave a damn and I think that came across.
A small note about Upscope
We hear that best in class companies convert 25% of their new sign ups.
You may have great features.
You also have confused users who don't understand those great features.
You have no time to set up screen sharing with each of them because it takes time to set up, it's blurry and involves swapping from one side to the other.
When they need help to understand your great features, it's a chance to amaze.
Upscope is "One click, see their screen and highlight where to go. Done!"
You co-browse together on the user's screen instantly and they enjoy it.