Product Modeling in B2B SaaS
TL;DR: When running product initiatives in B2B SaaS, it is important to have a model that connects this initiative to your overall company metrics. Without it, you’ll have a harder time convincing others that it is important.
What’s a model
Models provide an amazing understanding of the world. That’s why we create models for phenomena like the weather.
A model is a representation of a system. This is used:
- To understand
- To predict
In the case of our weather model, we use equations to represent the atmosphere in order to forecast the weather, i.e. to predict.
Product models
A product model attempts to represent if your product efforts are creating impact. As part of a normal product strategy, you’d like to connect your product efforts with the company strategy. A company strategy often has commercial metrics — such as growth and retention. A product strategy often has goals for users doing something — for example, engagement or achieving specific results. A product model should ideally connect the company strategy and the product strategy. And your model should connect the initiative to the product or company strategy.
Example: A simple model for a product initiative for Facebook could be modelled like this:
Facebook Friends → More Content → Spend more time → Exposed to more ads → Revenue
We will present it in a table format:
A model is built upon assumptions. This model assumes that getting more friends on Facebook will (in the end) increase revenue.
Top-Level Product Metric: Measuring Success
It’s very helpful to have an overall product metric that connects product efforts to company success as part of your product strategy. In our Facebook example, the overall product metrics could be the average time spent per user or weekly active users.
A North Star Metric is generally a good idea as a top-level product metric, but for B2B SaaS — where product-market fit in the first stages is so essential — I’d recommend Mark Roberge’s leading indicator of retention.
Leading Indicator of Retention
While you can discuss many definitions of product-market-fit, I find that retention truly is a measurement of product-market-fit: If customers stay customers, you have proof that they get value (they can’t easily get better somewhere else).
But retention in B2B SaaS is a lagging measurement. By “lagging,” we mean that we can measure other things more quickly to predict retention:
It is possible for companies to find leading measurements of customer retention. A few common examples of leading indicators of retention:
- Slack: The customer sends 2,000+ team messages (in the first 30 days)
- HubSpot: Customers use 5 features (out of 25 features) in the platform (within 60 days)
The leading indicator of retention connects an overall product metric with a business outcome:
You can then create product initiatives to make sure more customers achieve the leading indicator of retention.
Next-Level Metrics: Stable metrics
In a product model in a B2B SaaS product strategy, the next level of the model should be high-level metrics that are stable and impact the top-level metric. Stable in this context means a metric that doesn’t change for at least a year. I’ve seen countless companies where product teams get new OKRs each quarter with no connection to higher-up product metrics. The result is often that nobody really cares about the individual team goals, and there is no real connection to the overall company strategy. This can be confusing for a team and doesn’t show the value of the team’s work.
Having stable next-level product metrics creates the stability to work aligned towards goals. It won’t be perfect. It won’t measure everything (so don’t be religious), but it gives the stability to do real product discovery and experiments.
Example: A Product Model for Trustpilot
Let’s imagine that you were the new CPO of Trustpilot.
Trustpilot collects reviews about companies. It’s a service for users to read reviews about online companies and for online companies to collect and use these reviews to sell more. Some key features of Trustpilot are:
- Google Selling Rating: Get a star rating next to your link in Google Search:
- Trustboxes: Show off reviews on your website for social proof
- Reviews in Social Media Posts
As a CPO of Trustpilot, you could make a top-level metric for the product efforts first:
This top-level metric measures the value customer get from the product, and the user has to actively do something to get there.
Your next job would be to decide on your next-level product metrics and connect them to the top-level metric:
In this example, as a CPO, you chose to have metrics for both the consumer side (acquiring reviews) and the business side (utilizing reviews). For each next-level metric, you can work on product initiatives:
When a team works on an initiative, they should begin by constructing the model. Let’s imagine you set out to build a Shopify module, then you would initially build a model like this:
The team can clearly see how their work (building the Shopify module) contributes to an overall objective. With a model, the team is also forced to think through the entire lifecycle. It’s not just about building the module. Customers also need to install it, and this aspect needs to be considered in the work. I’ve written another blog post about how you measure in the different stages of the model.
Closing thoughts:
- If you run a B2B SaaS business, and you don’t have a product model for what you do, your only model is most likely sales numbers, and it is easy to push items on the roadmap based on what your sales department suggests.
- The accuracy dilemma: Models are not always a perfect or highly accurate representation of the world. You’ve probably experienced rain on a day it wasn’t predicted. Sometimes the model is wrong, but it’s often better than no model. All models are wrong. Some are useful.
- Avoid getting lost in the pursuit of the perfect model. It should be useful, rather than a precise 1:1 representation of the world. 1:1 maps tend to be impractical due to their excessive size :-).
