Survey-based Experiments


With Gradient's experimental design services, you can run tests with precision, not intuition.


We've seen it all too often — you run a survey, get an "interesting" dataset, spend weeks poring over crosstabs, get lost in the deluge, and exit the engagement as confused as ever about what decision you should make.

With Gradient's experimental design services, we do that work up front, and design a research study that will deliver definitive answers about what you should do next. Here's how we do it:

  • Structure the problem: gather perspectives from stakeholders, dig through past research, and design an overall objective that navigates tradeoffs
  • Build an objective function: this could be a well-designed survey scale, a revenue-maximizing concept in a conjoint, or almost anything that we can think of that ties directly to your KPIs
  • Design an experiment: this is where we get into the nitty gritty of power simulations, identify the number of possible treatments and interactions, the sample size we'll need for each level of confidence, and so on
  • Gather data: Gradient works with sample providers up and down the range of cost and qualilty
  • Run the analysis: here's where we run the appropriate statistical model to identify the best performing options
  • Optimize: Sometimes this is as simple as selecting the single best performing option; sometimes it means using state-of-the-art discrete optimization routines to search through a restricted space of trillions of combinations. Which one is yours?

Identify the best message

Confused about which message you should lead with? Disagreement within the team? Too many options and variations to choose between? Run an experiment. Gradient will show your message options to treatment and control groups, and measure their response. We'll identify the message that has the best chance of being your best performer.


Develop a data-inspired value proposition

Not sure which of your customers' problems you should tackle first? Which is the biggest value driver? Which is the biggest pain point? Use the tools the best product management orgs in the world use to build their product roadmaps: use elicitation techniques that develop a unique, statistically robust ordering of problem statements.


Find the optimal pricing and bundling for your service

Launching a product/service and not sure how to price it? Lost in the possibilities of which features should be bundled and which should be add-ons? Confused if you should offer membership, package, or pay-as-you-go options? What the pricing delta should be between variations to drive maximum price discrimination? In situations like these, there are often literally trillions of options (see above) to choose between. With a conjoint experiment and optimization routines, Gradient can cut through the noise and find the best way forward.