Pricing in B2B SaaS is something I’ve talked about before – specifically how we discovered that Freemium in SaaS can be deadly for an early stage startup searching for product-market fit.
Luckily (for Trakio!), I’ve evolved as a CEO and data analyst. Gone are those early days of jumping out of the gates with ‘gut instinct’ decisions on what seems “about right based on what others are doing!”.
When we started work on the relaunch of Trakio’s newest product, Metrix (aka a “pivot”), one of the areas of the business I knew needed much more work, and more strategic consideration than I’d given it before, was our pricing model.
This is a guide to the thought process and analysis that has gone into our new pricing model for our launch.
I knew I needed to look again into what makes a good value based metric for pricing. I needed to re-examine Freemium and it’s failings, and look into how larger and more expensive B2B SaaS products commanded such high prices against smaller, cheaper competitors.
In Trakio’s history, we’ve had 3 different business models:
- Model 1: Our first product used Freemium (a free plan, plus a free trial on paid plans). Premium upgrade packages were based on app/website usage volumes and were designed to simply undercut every analytics product on the market.
- Model 2: Our second product started at $2,000 /mo (with quarterly or annual payments only) and was a fully managed SaaS+Service hybrid. Pricing increased on customer volume
- Model 3: Our third product (Metrix) is priced per user/seat, with a very low starting price but scales up quickly for larger businesses. There is no free trial or free tier, but there is a refund guarantee.
Model 1: The Hard Thing About Freemium For B2B SaaS
Freemium can be a great growth vehicle for B2C and B2B SaaS. However, as I discussed in great detail in an earlier post, most successful B2B SaaS companies using Freemium today did not start out as a Freemium company.
Most of the successful B2B SaaS businesses with free tiers (and even free trials) today, actually launched their product using paid-only options, and added Freemium only once they had significant product-market fit and/or significant venture capital to subsidise losses.
- Mailchimp have a great history lesson on their move to freemium.
- Jason Fried from Basecamp has admitted that nearly all of their revenue comes from paid plans (about 6:00 into the interview), and those users pretty much all started on paid plans upfront (i.e. they didn’t upgrade from a free plan).
- CrazyEgg killed their Free plan almost immediately after launching, and doubled their revenue in that same month.
- BidSketch killed their free plan and saw a 10x increase in paying customer conversions in the first month
There are various ways Freemium can kill an early SaaS product, such as high server costs, reducing the perceived value of the product or attracting the wrong customer segment.
Our own personal issue with Freemium was that we were still 100% in customer development and we needed to get great product feedback. We wanted to fine-tune and tweak the product so that people would pay us a lot of money for it to solve their pain.
However, all of our free users were suggesting “cool” features, and were drowning out the voices of our customers who were suggesting “valuable” features. Our free users were pretty much all tiny projects, unfunded startups, or simply businesses that had poor business models and were unlikely to ever have budget (or need) for good analytics software.
Free Users Help You To Build A Cool Product. Paying Customers Help You To Build A Valuable Product.
Model 2: The Hard Thing About Pricing Enterprise B2B SaaS
Our second product pivot, which we started as we landed in San Francisco and joined an Enterprise focussed B2B accelerator, was mostly a pricing and service pivot. At first.
We chose a price point of $2,000 /mo based 90% on gut and 10% because that’s what we’d “heard our competition charged”.
Once we got tighter into the Silicon Valley community, we soon learned that these competitors only used $2,000 /mo as a low anchor point, but aimed for a $6k-$10k /mo deal size.
Or the ones who weren’t going for larger deals, and were focussing on the $1,000 – $2,000 /mo customers, were massively struggling with huge customer acquisitions costs, churn rates approaching 50% and/or were burning through millions in venture capital (while only making small growth compared to typical B2B SaaS benchmarks).
The reason is: $1,000 – $2,000 /mo is a very difficult price point to make true enterprise sales work, especially considering we had a complex product that required a high-touch service.
Many senior SaaS advisors (product managers, CEO’s, investors) warned us that $2k would be “Startup Graveyard” for our very complex (i.e. ‘analytics’) product, but I carried on kicking doors down anyway!
Soon enough, a pipeline started to fill up, and demo requests became pilot proposals! Then pilot proposals became feature requests. Feature requests became messy and time-consuming, the pilot pipeline all got a bit muddy and things started to slow down.
And then we realised that starting an enterprise B2B SaaS startup with any less than 8 people and $1million was crazy town!!
Model 3: The Benefits Of Per Seat Pricing For Metrix
Benefit #1: Very Low Friction
Metrix has an initial pricing point of $10 per month per seat. Which means for $10 per month, an customer success manager at a subscription company can connect 2 or 3 of the SaaS products they use (perhaps just the tools that the CSM has access to) and load all of that data into the Trakio platform, and then download the Metrix app from the App store and start playing with their data. All in a matter of minutes.
The low payment is small enough for a personal or a team credit card without any approval. There are no budgets to get approved, not pilot processes, no messy internal processes… even when a company commits to a free trial, many large companies still require employees to get preliminary approval of what the final pricing would be (something we discovered when ‘selling’ free pilots on our Enterprise platform!)
Benefit #2: Scales with Larger Companies
Any good value-based pricing metric should scale with larger accounts. Per-seat pricing for Metrix means that as we sell into larger and larger companies with more teams, we can capture more revenue as they deploy larger rollouts of Metrix.
This is a tactic used to great effect by Slack – while only a few bucks per user, when large 2,000 employee companies decide to do a full rollout, the numbers start to look really healthy for Slack!
More mature companies pretty much always learn that they need to actually increase their pricing for large accounts so that it scales higher than linearly. Initially, this seems counter-intuitive to the customer: more seats should mean some type of volume discount?
However, servicing a 2,000 employee company is usually more expensive than servicing 10 x 200 employee companies. This is where paid upgrade features come into play such as an enterprise-grade SLA, or paying for the dedicated server resources availability.
For Metrix, we haven’t put too much concern into this yet to avoid risking complicating things before it’s necessary.
Benefit #3: Land and Expand
Metrix is specifically designed to solve the problems associated with data silos within companies. This means that it becomes exponentially more valuable as more teams and departments in a company use it.
However, trying to sell an entire organisation at once is extremely difficult. Instead, our pricing model allows one department (perhaps the one with the most burning desire for Metrix, or the most data-savvy) to use the product for a while first. Then once we’ve established some evangelists in the team and can build a good internal case study, we can work with that team to get other departments onboarded.
It seems strange to talk about an “analytics product” in this way, but remember that Metrix has been built for anyone to use. It’s a self-service tool to empower everyone in the company to make data-driven decisions, and so we’ve looked to examples in collaboration tools rather than vertically-focussed data science tools.
Benefit #4: Increased Buy-In = Stickiness
Having your internal champion leave a company sucks. I know because we had it happen to us with one of our largest (and probably most famous) customer in our early Beta. This can happen where one or two people in the company have wanted an expensive, but specialised, tool for their role. They completed all the due diligence and fought for the budget to buy-in the software – why wouldn’t they if it made their job a lot easier each day.
But if this person leaves because their position is dissolved (in our case due to an acquisition), it can be extremely difficult to smoothly transition to another champion within the company.
We learned the importance of team buy in from this painful lesson pretty early, and it’s a lesson I preach again and again to B2B SaaS owners in our training courses and through calls on Clarity.
With Metrix, we’ve amplified this and gone beyond simply having the full department onboard – we want the entire company onboard! There are only a few categories of apps that every employee in a company uses (chat, project management, email, maybe BI, maybe CRM…) and we’d like Metrix to be one of them.
Benefit #5: Only Pay More For Higher Value
Most analytics providers deal in charging on data volumes. This means companies or businesses who are more data-savvy and have more data, are penalised for it. While in some cases, more data can mean you are more likely to get higher value from an analytics tool and therefore a higher price is fair, there are situations whee this doesn’t hold true.
For example, if only 1 or 2 people in your company are making data-driven decisions on your data (probably the product management team). When the marketing team do a great job and start generating a lot of extra downloads on your app, you’ll see increased charges in your analytics provider, but it’s unlikely that all of this extra data will bring any significant value to those 2 people in the product team. The company pays more, but doesn’t get extra value from their data.
In the case of Metrix, we only charge more when the size of the team leveraging that data increases.
Model 3: Risks Of Per Seat Pricing For Metrix
Risk #1: Only A Handful Of Users In A Large Company
The immediate risk that came up for us when we chose this pricing strategy was what if only a handful of employees in a large company signup, but have a LOT of data? We’ll lose money on server costs, lose a lot of missed revenue and maybe face the issue of not been seen as a ’serious’ vendor to that company.
Of these risks, the most immediate is the server costs and potential to have crippled performance. For this reason, we’ve added a “sanity cap” onto our product. It’s extremely generous and based off of the 300+ SaaS companies who have sent large volumes of data to us in the past.
This limit captures the sorts of companies who would want to open a dialogue before signing up anyway, or at the very least, within days of a free trial. By having this conversation first and moving them onto our ‘Business’ Hub tiers, we physically put their account on a separate database server cluster which is capable of delivering the same snappy performance they’d expect, even when they throw 1B monthly records at it!
The ‘middle ground’ scenario is a company who aren’t quite ready to spend an extra $1,000+ /mo on their business upgrade, but have quite a lot of data, but are still wanting to just test out Trakio. In this case, it’s simply our job to ensure we build a product that genuinely increases in value as the number of team members increases. We’ll make a loss while the customer has 1-10 seats, but after that we’ll move into profitable territory.
Risk #2: Getting Less Serious (High Churn) Customers
A product with a $10 /mo price tag is going to attract a lot of ‘casual’ users. And casual users churn quickly, don’t respond to feedback emails and are very unlikely to participate in case studies, webinars etc.
To combat this, we’ve added the payment information field upfront on signup, before the user has a chance to get into the app and start importing data.
While this will cost us a lot on signup numbers, our experience has shown us that there is a huge difference between a user paying $0 (Free trial or Free plan) and having to pay just $1. By ensuring that the user has a burning need… or at least one worth spending $10 upfront on, we can at least filter out some of the less serious customers.
Beyond that, we just need to handle our account management well and setup different customer success processes for users who stay on 1-5 seats vs. accounts who quickly upgrade to 25-200 seats.
[In our experience, nearly all of the accounts on our lowest paid tiers who churned actually churned because their project/startup died]
There is a lot of criticism for per-seat pricing in SaaS… and I think I’ve read pretty much all of it. In summary, most of this criticism is from people who still have unproven business models or who are trying to simply differentiate themselves from an incumbent with a pricing model that’s cheaper to smaller startups (i.e. offering a ‘5 Agent tier’ for cheaper than the price of 5 x agents at their competitor).
And to be clear, when I analysed all of the successful startups that used a per-seat model and broke down their value metric, the model fitted well. Likewise, when I broke down all of the successful B2B SaaS startups with usage based value metrics, they also seemed to fit the model of a good value based pricing metric.
This post isn’t about saying “all B2B SaaS should be per-seat or per-agent”. There are many lessons in this post, mostly learned from our failures. What we learned is that your pricing should come from a thorough research into the psychology of your customer, how their organisation is made up, how you intend to sell to them, how you intend to grow with them, how you intend to service them….
…. do not price your product to undercut a competitor. Do not price your product based on what an ‘apparently successful’ competitor in YC is doing. Until companies grow to IPO stages (or you start building inside connections and can find out their actual revenue numbers!) it’s going to be very hard to know if their pricing model is actually working, or not.
And for fuck sake, do not add a free plan until you have at least $1m ARR!