Tell us about how you went from defining your North Star to actually calculating it, and turning it into a metric that you can follow up. What steps were involved? What key considerations did you make along the way?

This was quite a lengthy process. After defining the North Star as “Minimise the time spent to acquire goods and services to do my job”, we started working on the formula in October 2023. The first iteration of it was then deployed in April this year (2024). 

From our point of view in the data department, the most important thing was to make the metric easy to understand and explain, because if we can understand it, we can influence it and improve it.

In order to do that, we scheduled multiple sessions with key stakeholders where we discussed how we define “goods and services” and how our customers spend time on acquiring them. We started with open discussions, letting people express their opinions and brainstorm ideas. After that we narrowed it down to multiple options, each with pros and cons, and discussed them with the stakeholders, finally coming up with a clear definition of what we want to measure. During these discussions we made sure that we actually have the data we are talking about and that it would be possible to obtain with the resources we have now.

We also had to keep in mind that it is not all about the end result but also about the journey. In other words, it was important for us to measure if our product is being used as intended and optimised, not just that it was being used. To do that, we created strict definitions of what we consider successful product usage, so we can notice in the North Star if our customers are not using the product optimally. 

Once that was all set, we went to work by gathering the first batches of data, seeing if the values we came up with made sense and was something we could work with. On some calculations we had to go back to the stakeholders and discuss alternatives, but in general it was a smooth process.

What specific types of data did you find necessary for calculating your North Star metric, and how did you go about collecting this data?

We decided to go for an average time spent per successful item, so our calculation basically consists of two parts: time spent in our product and successful items created.

We are getting the Item data straight from our product’s database, where we have set up daily exports to the VDL, which allows us to work stress free with our product data. 

The Time spent data was slightly more complex. We have both a mobile application and a web application. Initially, we were aiming to implement a solution that would measure engagement time in both applications, but after talking with our app developers, we learned that all that we need is already available in our Firebase Google Analytics database. As for the web app, we went with the Snowplow solution. This was already on our radar and something ProActive had tested out before. Together with the Product Analytics team we were able to revive this project and get it up and running with ease.

Now with usage data already in VDL, Snowplow data being in the Snowplow BigQuery project, and the Google Analytics data exported to another BigQuery project, we had the foundation set up. To combine it all together, we used a service account that had read access to everything and our DBT project. It was quite easy to set up from there. We aggregated all the data per customer, module and day separately, and then joined together in a final table. Our final table is an incremental table, as we wanted to keep the historic data as it was. All this is run daily and updates for the last 3 days, because sometimes we have some records coming in late.

‍

How do you calculate your North Star? Did you need to adjust this calculation along the way? Take us through the details of your North Star Formula! 

The calculation is fairly simple, as we also aimed to make it so. It is an average of averages. Our product has different modules, each with slightly different, yet similar use cases.

Some modules are used more actively than others, but for us they are all equally important, so to make sure that each module contributes to the North Star equally, we decided to calculate a “mini” North Star for each of them, then average it out as the end metric. This is our Formula:

‍

‍

Additionally, we defined a “General” module that is also part of the calculation. The General module includes all the time spent on our product that is not directly related to a specific module, like application management, and all the successful items.

We had other options involving the monetary value of items or spend as well, but decided to go with this formula as the value of spend itself doesn’t reflect on the acquired goods and services as well as a successful item does.

However, after just a few months of the initial North Star metric being live, we decided to remove the Contract module from the formula and instead add the engagement time of it to the General module. This was due to the fact that success of contracts are very hard to measure with items, it’s rather a combination of time and invoices linked to the contract, which did not fit the overall formula.

‍

Do you have any plans to improve your North Star Formula in the future?

Yes, we plan to improve the granularity of our data and add more dimensions to it. We already have multiple requests from our colleagues who want to dive deeper into it. A few examples we can name are:

• Being able to split the North Star value on a workflow, like the encoding and approval part of our invoicing process
• North Star from the point of view of invoice/order suppliers, to see if we can optimise the process for the biggest ones
  
  

How do you visualise your North Star Metric and make it available to all of ProActive?

First of all, we made sure to integrate it into a management dashboard that our leadership uses weekly.

That dashboard is based in Tableau. The same data was also added to our OKR dashboard.

In addition, we created a separate Tableau dashboard solely about the North Star with different drill down opportunities.

These dashboards are shown on screens around the office and shared every quarterly kickoff.

We have also set up a customer maturity model, which essentially is a Google Sheet containing all the most important product usage metrics for every customer on different time intervals (30, 60 and 90 days). We added the North Star values here as well.

Lastly, the North Star data is also available in its raw form in the VDL so that any ProActivian interested can access it. 

‍

You have made an impressive improvement to your North Star Metric, not least this past month! Have you made any key improvements to the product that have actively had an impact on your North Star?

Yes, we have made several improvements that we believe have had an impact on our North Star Metric. Firstly, we have created a specific order request for services to make it easier and less time-consuming for our users to submit these types of requests.

We have also introduced an onboarding flow in the app for users who get a payment card assigned, as this should help minimise the time it takes to understand how to use it for online and offline payments.

Going forward we are focusing on the encoding of invoices. I spoke to a developer today who said that the North Star helps them a lot in solution design. They were discussing "setting default values" to minimise time in encoding, but someone remarked that the application management involved in this is also counted in our North Star and should therefore be taken into account. This is exactly the outcome from implementing a North Star that we are looking for if you ask me.

What advice would you give to other Visma products looking to calculate or visualise their North Star Metric effectively, to create the same kind of engagement that you have in ProActive?

• Involve stakeholders from all parts of your organisation, because a North Star is a global value that should be relatable to everyone!
• Keeping it simple will go a long way. If you can explain it to your co-workers or customers with a couple of sentences, it will be so much easier for everyone involved to track it and be engaged in improving it.

• Don’t try to make the perfect version on the first try. Obviously, you would like to keep it the same over long periods of time, but having a few fixes or changes along the way won’t do harm. If anything, it will create more discourse and awareness around the metric.
• Use the fact that you are part of Visma Group. There are many opportunities to learn, like this article, like the Product Discovery & Analytics team with their expertise and much more. You just need to reach out!

‍

‍