Thanks for walking through all of this! Really cool to see the progression in detail.
I'm curious how you realized connectors were not only the best determinate for value, but also that willingness to pay differed among them? Was it based on intuition from talking to users and sales calls? Did you uncover in max diff survey?
And when figuring out each of these different version, did you launch each iteration of pricing as an a/b test?
Re: Connectors – a lot was learned qualitatively via direct conversation with leads. Also, we could see in usage and data that customers with more than one connector were typically more engaged/better customers for us. On the willingness to pay front, a lot is predetermined by how much the customer is already paying for the tools they are connecting to. For example, Salesforce is usually used by larger companies and is expensive. We also incur higher costs to run specific connectors, which informs how much we charge for them too.
Re: A/B testing - I've personally never A/B tested pricing. In past roles, we've either done a lookback analysis (aware of the caveats and trade-offs), i.e., comparing previous cohorts (old pricing) to newer cohorts (new pricing), or tested in specific geographies before committing to a global rollout—also with its caveats.
At Equals, we've simply changed for all new customers with each new version, knowing that at some point, we'll want to migrate customers on legacy plans to current pricing, ensuring they're treated fairly with grandfathering, etc.
The primary reason for never opting for an A/B test is the need to learn and act as quickly as possible. Split tests simply take longer, especially in B2B when you're at an earlier stage and your sample size is naturally smaller. Typically, you also need to wait at least a few months to see how cohorts on newer pricing age to understand the impact of your changes on expansion, contraction, and churn over time. e.g., ACVs on new pricing might look great, but perhaps they don't expand as well and churn faster.
Thanks for taking the time to give such a detailed response. That is helpful! I really like the point about judging willingness to pay for connectors based on the price of the connectors themselves. Super smart and not something I thought of.
Re: testing — At my company today we are more “prosumer” and so operate closer to a consumer modal. And while we have decent sample sizes, we still typically have to wait months to fully understand impact of our split tests on pricing — we also have complexity of gradual user journey were adoption takes time based on real world behavior/constraints. So I 100% understand that challenge being greater with true B2B.
Really interesting perspective. Thanks again for taking the time to respond :)
I guess I should have mentioned that I have A/B tested pricing page design in the past, but only when the underlying model for each variant is the same.
Great post matt - how did you decide on that shift from self-seve startup focus to sales-led (smb/mid-market focus) ? Did you have them running always side-by-side or did you start with a sales - first Motion and was self-serve an experiment which didn't work out or bring enough conviction?
We started gated when the product first launched. We then opened up self-serve and about 3 months ago went sales-only. Self-serve wasn’t an experiment so much as our product, positioning, and pricing were not optimized to make it a successful motion for us at this stage. We’re hopeful we can make it work in the future though.
This is great, thanks Matt! Is it monthly contract or annual? If annual, what do you do when folks want to add another integration? If monthly, do you worry about churn?
Thank you, Stephen. Annual contracts only at this stage. As for adding an additional connector mid-contract, there is some nuance but the simple answer is we effectively issue an invoice to cover the additional cost.
Awesome, thanks! We were considering whether to charge connectors individually or as a bundle. I was worried individual would be more complicated (triggers call with sales, gotta pay again, customer thinks "we can get by for now") whereas bundle is just "we already paid for it, let's turn it on". Sounds like it may be more simple than that. And obvious downside to bundle is folks are pissed if they pay monthly for 5 connectors when they only need 2-3 8/12 months or something
Maybe, maybe not. We’ve talked about selling bundles and might one day. There’s some
Engineering effort to make it work (well).
Also, I’d wager not all connectors are equals (no pun intended). For us, given the use cases we build for, it’s uncommon that a connector is only needed ad-hoc.
Take ARR reporting for example - you want the Stripe connector ON every month. Same goes for Xero or Quickbooks if you’re doing your P&L reporting.
So while annual works for us in this context, I can see why it may not work for others.
Thanks for walking through all of this! Really cool to see the progression in detail.
I'm curious how you realized connectors were not only the best determinate for value, but also that willingness to pay differed among them? Was it based on intuition from talking to users and sales calls? Did you uncover in max diff survey?
And when figuring out each of these different version, did you launch each iteration of pricing as an a/b test?
Thanks, Luciano!
Great questions.
Re: Connectors – a lot was learned qualitatively via direct conversation with leads. Also, we could see in usage and data that customers with more than one connector were typically more engaged/better customers for us. On the willingness to pay front, a lot is predetermined by how much the customer is already paying for the tools they are connecting to. For example, Salesforce is usually used by larger companies and is expensive. We also incur higher costs to run specific connectors, which informs how much we charge for them too.
Re: A/B testing - I've personally never A/B tested pricing. In past roles, we've either done a lookback analysis (aware of the caveats and trade-offs), i.e., comparing previous cohorts (old pricing) to newer cohorts (new pricing), or tested in specific geographies before committing to a global rollout—also with its caveats.
At Equals, we've simply changed for all new customers with each new version, knowing that at some point, we'll want to migrate customers on legacy plans to current pricing, ensuring they're treated fairly with grandfathering, etc.
The primary reason for never opting for an A/B test is the need to learn and act as quickly as possible. Split tests simply take longer, especially in B2B when you're at an earlier stage and your sample size is naturally smaller. Typically, you also need to wait at least a few months to see how cohorts on newer pricing age to understand the impact of your changes on expansion, contraction, and churn over time. e.g., ACVs on new pricing might look great, but perhaps they don't expand as well and churn faster.
Hope that helps!
Thanks for taking the time to give such a detailed response. That is helpful! I really like the point about judging willingness to pay for connectors based on the price of the connectors themselves. Super smart and not something I thought of.
Re: testing — At my company today we are more “prosumer” and so operate closer to a consumer modal. And while we have decent sample sizes, we still typically have to wait months to fully understand impact of our split tests on pricing — we also have complexity of gradual user journey were adoption takes time based on real world behavior/constraints. So I 100% understand that challenge being greater with true B2B.
Really interesting perspective. Thanks again for taking the time to respond :)
No worries. Glad to help.
I guess I should have mentioned that I have A/B tested pricing page design in the past, but only when the underlying model for each variant is the same.
Great post matt - how did you decide on that shift from self-seve startup focus to sales-led (smb/mid-market focus) ? Did you have them running always side-by-side or did you start with a sales - first Motion and was self-serve an experiment which didn't work out or bring enough conviction?
We started gated when the product first launched. We then opened up self-serve and about 3 months ago went sales-only. Self-serve wasn’t an experiment so much as our product, positioning, and pricing were not optimized to make it a successful motion for us at this stage. We’re hopeful we can make it work in the future though.
This is great, thanks Matt! Is it monthly contract or annual? If annual, what do you do when folks want to add another integration? If monthly, do you worry about churn?
Thank you, Stephen. Annual contracts only at this stage. As for adding an additional connector mid-contract, there is some nuance but the simple answer is we effectively issue an invoice to cover the additional cost.
Awesome, thanks! We were considering whether to charge connectors individually or as a bundle. I was worried individual would be more complicated (triggers call with sales, gotta pay again, customer thinks "we can get by for now") whereas bundle is just "we already paid for it, let's turn it on". Sounds like it may be more simple than that. And obvious downside to bundle is folks are pissed if they pay monthly for 5 connectors when they only need 2-3 8/12 months or something
Maybe, maybe not. We’ve talked about selling bundles and might one day. There’s some
Engineering effort to make it work (well).
Also, I’d wager not all connectors are equals (no pun intended). For us, given the use cases we build for, it’s uncommon that a connector is only needed ad-hoc.
Take ARR reporting for example - you want the Stripe connector ON every month. Same goes for Xero or Quickbooks if you’re doing your P&L reporting.
So while annual works for us in this context, I can see why it may not work for others.