A software company is launching a new analytics add-on for its existing customers. How would you set the price?

Before I pick a method, what's the goal for this add-on? Is it meant to grow revenue per customer, or mainly to keep customers from leaving? I'd come at the price from three angles. The first is value to the customer. I'd find out what customers are willing to pay, probably with a survey. The second is alternatives. I'd look at what competitors charge for something similar. That gives a reference point customers will compare us to. The third is cost. I'd only use it as a floor, to make sure we don't lose money on it. Then there's how to charge, which matters as much as the number. It could be a flat fee, per user, or bundled into a higher tier. I'd look at how the main product is priced. I'd start with the customer survey, then test a price. Does that sound reasonable?

Before I pick a method, what's the goal for this add-on? Is it meant to grow revenue per customer, or mainly to keep customers from leaving? I'll assume the goal is more revenue per customer without hurting renewals. I'd come at the price from three angles. They don't carry equal weight for software, so I'll take them in order of importance. The first is value to the customer. I'd find out what customers are willing to pay, probably with a survey. What I'm after is what the add-on saves or earns them. If it replaces ten hours of analyst time a month, or a $500 a month tool, that sets the ceiling. I'd expect that to differ a lot between small customers and enterprise ones, so I'd look at each segment separately. The second is alternatives. I'd look at what competitors charge for something similar. That gives a reference point customers will compare us to. It includes standalone analytics tools, and it includes doing nothing, which is usually the real competitor for an add-on. The third is cost. I'd only use it as a floor, to make sure we don't lose money on it. Serving one more customer costs almost nothing in software, so cost-plus would badly underprice it. Then there's how to charge, which matters as much as the number. It could be a flat fee, per user, or bundled into a higher tier. I'd look at how the main product is priced. If that's per seat, I'd test a per-seat add-on at two price points, something like 20% and 30% of the base price, with a sample of customers. My hypothesis is that it should be a paid add-on priced on value, with a bundle for enterprise. First I'd ask for usage data on who needs analytics most. Do we have that?