Pricing, Packaging, and Compute Power
Pricing and packaging flexibility drives subscription growth. Warning: this blog is going to be hefty, and I’ll be referencing a talk from a subscription conference, a competitive blog, and explore different pricing options for the current space I’m in.
My former Codesion CEO, Guy Marion, is now the Sr Director of Online and Customer Acquisition at Zendesk. He has been focusing on customer acquisition, engagement, and lifecycle growth. At Zuora’s Subscribed ‘13 Conference, Guy shared five secrets of growing a subscription business, and driving customer acquisition through pricing and packaging. You can see the full preso at http://www.youtube.com/watch?v=AEwQgrNbo5E
Secret 1: Start simple with a low friction offering. SaaS is focused less on the high up front ACV/TCV, and more on the long term value of a customer. Land and expand. Build a community base early on, with a self service approach. Offering service via a freemium model or trial is up to you, but make sure it’s easy for customers to sign up and navigate. The website should enable your customers to get enough information online and understand your basic differentiators.
Secret 2: De-risk your pricing launch by testing the waters first. Test quickly and iterate and find established product market fit. As you mature you need to focus growth and preservation of your base. Continuing the first secret, the key with SaaS is maintaining long relationships, so validate the new pricing with AB testing, roll out to new customers as you see fit, and then you can eventually grandfather existing customers.
Secret 3: Evolve multiple editions to drive revenue while staying competitive. CollabNet Cloud had different iterations varying from Silver, Gold, Platinum, to Developer, Professional, and Enterprise. Make sure your pricing paves a simple and transparent growth path for customers.
Secret 4: Build a repeatable low touch sales model to reduced customer acquisition costs and focus your upstream sales efforts. Having a low touch sales model forces you to institutionalize best practices and builds a blueprint for what makes a customer successful. Understand your customer journey by enabling feedback mechanisms, tracking metrics, and knocking down hurdles that are preventing customers from on-boarding. Automate everything: Automate billing, invoicing, renewal process, keep systems aligned. The data in Zuora should be synced with your internal systems and Salesforce.
Secret 5: In SaaS, 90% of your MRR comes from existing customers, so price and package to upsell. Include upsell triggers in product messaging, email automation, and have proactive health checks and customer outreach. Have a customer success program with people who are incentives to keep your existing customers happily growing. Understand why customers churn, when they churn, and monitor churn risk to prevent further customer loss.
Our business model is:
B2B SaaS (multi-tenant)
Hybrid sales: Direct, self-serve and sales assisted business model, depending on customer needs.
A couple channels, mostly AWS and Akamai
Within our log management industry, there are competitors varying from enterprise on-prem, developer SaaS, enterprise SaaS, and open source. Kord Campbell, former Splunk employee and Loggly founder, wrote http://www.tinyprobe.com/blog/wacky-log-pricing.
Ideally, we’d all price similar to Google, which is based on storage and searching. Kord gives a “compiled list of all the the other company’s features/knobs/limits/crap you have to wade through to figure out how much sending them your logs is going to cost you and how long you’ll have access to your data:”
GB storage pricing (less the full text searchable index)
GB sent pricing
variable retention time limits (7, 15, 30, infinity days)
max storage size retention limits
daily volume limits
monthly volume limits
indexing hard limits
overage fees (daily/monthly)
extended support options
Sumo Logic received favorable criticism, granted that it takes a while to click through several tabs and finally find the pricing site…Whew! We’re in the process of hiring a marketing person who’s had more experience with product marketing, online-developer marketing, and quantifying successful marketing channels.
Pricing levers depends on the product and service. Pricing also depends on customer perspective, pricing models they’re used to, and also competitive pricing models to make it easy to understand and compare. Pricing and packaging is critical - it can convince your customers that investing more time in evaluating or jumping right in a purchasing. Or it can be confusing enough to deter from taking the next step.
Let’s look at Kord’s idea. What if log management services adopted pricing simply based on storage & searches?
While storage is easy to measure, I think measuring searches is complicated. First of all, searches vary on complexity, and there are different types of customers who expect certain performance. A user may be using the service for business critical operations, is investing in your vendor product to support their own product/offering, and may have SLAs with their own customers. Another user may be using the service for a personal project, smaller implementation, or testing purposes, and service performance does not impact their bottom line.
In this case, while you can provide a pricing model for storage and number of searches, there needs to be a way to guarantee levels of service performance. One way you can identify levels of performance is by pricing based on search complexity or compute power. Sumo Logic is awesome because they guarantee search performance and have identified "acceptable" rate of search performance.
Search metrics depend on several variables:
Selectivity: Selectivity is defined as number of hits within a total number of messages contained within the search scope. For example, if there are 1 million logs in an hour, 100 of which have word “error” in it. If you query for all errors in the last hour, the selectivity of the query is 1 in 10,000.
Time Range: Time range is the duration over with the query was run. For example, if you run a query over the last hour, time range is 1 hour long.
Search Completion Time: Time for the search to complete and return all results that match the search criteria within the time range of that query. SLAs and pricing levels based on compute power would be ideal since you’d really be paying for what you use. However, I’m not sure how scalable it’d be and how the public would react to this kind of “expensive search” pricing model, especially since compute power hasn’t been introduced before, and most log management vendors are pricing based off “total volume"…Thoughts?