What was done?

A new view introduced to see all and edit your carbon inventory.  It replaced a view where clients, climate strategy advisors and data specialists spend the bulk of their time in the Normative app, outside of excel-sheets.

Meet the team

1 Product manager
1 Product Designer (me!)
3 Front-end engineers
2 back-end engineers

Let's start with the product vision

Normative will always need to support easy measurements for our customers. To offload the burden of the complexity in Carbon accounting, our aim is to ensure automated solutions wherever and whenever it’s possible.
However - looking at the current reality and where the knowledge lies (with our customers) we need to support a way for manual data refinement.To get to a more accurate estimation of greenhouse gas emissions, customers need and want to provide activity data and supplier-specific data. Customers have first hand information about their carbon footprint and should be enabled to own the creation and refinement of their data to make it decision useful.

From this product vision, me and the PM defined some intial How might we-statements to guide our work forward.

How might we?

Shift manual work to our clients from Normative employees to scale the company?

Drive customers to "level up" their data quality?

Guide the user through this?

Pain points

1. No overview of your inventory

The current view the Supplier view does not show all the data you have uploaded, only the financial information. So kWh spent on electricity, heating or petrol/diesel spent on your vehicle fleet is not represented here. To see that information there are dedicated dashboards capturing all emissions but no place serves as an overview where we can enable edits across all green house gas categories.

2. Customers lack ownership

As soon as the customers uploads a data file into normative they loose ownership of the data, the file needs to be manually processed, validated and quality assured by Normative employees. If the customers want to edit the data they have to delete the file and upload a new one with the edits the want to make. Customers should be able to themselves perform manual actions (refinements) to ensure accurate measurements. And work with activity data to increase accuracy as a first step in a reduction plan.

3. High cost to serve

Decrease cost to serve for internal teams to be able to scale as a company to move away from manual uploads and other processes.

Jumping into discovery

Since me and the Product Manager had a good idea already about what the pain points were and what we needed to solve, I jumped into the discovery phase on how to solve the problems mentioned with the intent to build a prototype for user testing.

Discovery tools

First me and the Product Manager had a few sketch-sessions to bounce ideas and outline problems
User journey-map
Desk research on competitors and other SaaS-applications
Stress testing concepts with actual customer data in lo-fi wireframes.

User research

I wanted to test the hypothesis of one page to see all activities both transactions and non-transactions, named “Activities”. As well as to simplify the editing process to see if it would empower users to take ownership of their data after it has been uploaded to the platform. To influence users to refine their data to be more accurate I also tested out a accuracy indicator to see how users responded to it.

Testing approach


Internal data Service Analysts


Sustainability managers (not using Normative - to not bias them to how things “usually” work.)

Study goals

General impressions
How are the GHG groupings perceived. Impressions on the sorting, pagination and groups in general.

Edit and undo
Do they understand where to add data, how they would edit or remove data.

Excluding activities
Asking users to remove activities to avoid double counting the same data coming from different sources. (A big problem in Carbon accounting)

What is accuracy?
Testing out a new concept to highlight the accuracy of activities.

First round of testing

The first round of testing was performed on internal Data service analysts to get all major UX-problems sorted before the external testing round. It gave us some  insights about short-comings of the experience, a lack of guidance was raised as well as the new groupings of activities was welcomed as a nice feature.

Second round of testing

Second round of testing with external sustainibility managers not having used Normative.

Key takeaways

I wanted to test the hypothesis of one page to see all activities both transactions and non-transactions, named “Activities”. As well as to simplify the editing process to see if it would empower users to take ownership of their data after it has been uploaded to the platform. To influence users to refine their data to be more accurate I also tested out a accuracy indicator to see how users responded to it.

What to call this view?

I gave the users a selection of options to choose from and asked them which one represented the view the best: Accounting, Carbon Accounting, Activities, Inventory

👎 Accounting speaks to financial accounting but we also show non-financial information so accounting is not a fitting name.

👎 Carbon Accounting feels a bit too general as data collection is also a part of the process of carbon accounting. Some users said it sounded more official, but less intuitive.

👍 Activities was most positively received as a name however 2 non-users perceived activities as physical activities they do in their line of business example packing, bottling for a Brewery.

The amount column

The hypothesis was since only one value is used for the emission factor, if we have data like kg/litres we ignore the financial cost;
I was debating if these columns could be combined.

💡 This was more intuitive for DS-users since they understood only one value was needed for the calculations, however some thought if you saw kg for an activity then cost was missing and not that if it was provided and accessible in the details sidebar.

💡 Sustainability managers expressed the need to be able to weight cost against emissions since it’s something they always have to prove internally if something is both expensive AND carbon intensive they get more buy-in to replace to greener alternatives.

👎 Would also be harder to spot double counting in the cases where one activity has both activity data and financial cost and the other only financial cost.

The edit experience

How can we make it easier to understand how to edit?

👍 100% of users completed the task of editing when told it was possible on the page.

💡 The mental model of a few users that are more used to working in spreadsheets want to go to the data collection tab to upload new sheets with the desired edits and see the activities tab as read-only

💡 When told to add kWh to a certain supplier from the tab itself the majority clicked the 3-dot menu > edit to add it.

👍 In the old view, there was no clear affordance to show that the table rows were editable, adding a specific menu-item to edit instead of just relying on users to intuitively click the row helped the less tech-savvy users we interviewed.

Undo / reset

How would users undo the edits they have made?

👍 100% of users could complete the task of undoing their changes

👎 Internal and non-users both leaned towards to erase the values of the input fields in the edit modal instead of clicking “Restore button”.

👎 The “restore original values” button was overlooked by half of the users. After being shown the button some voiced they would want it more in relation to the edit they made.

👍 The undo-icon helped them to understand the button since the button text was a bit hard to decipher.

🚧 All users went back to the edit-modal to undo, so we have to think how to increase discoverability of the sidebar.

The accuracy concept

How would users undo the changes they have done?Hypothesis to see if indicating the accuracy level of the data would nudge users to provide more granular and help them to figure out what to do on this view.

👍 Generally users liked the concept and welcomed us highlighting where you might want to improve. A high score would be higher accuracy on the assumption made.

🚧 Not super clear what the accuracy is indicating, based on the tooltip text users only make the connection between activity data = high accuracy. We should also spell out that cost = low accuracy.

👍 Users found it easy to figure out what it is by hovering the tooltip

🚧 Users felt the accuracy was only connected to the amount column, if a very broad taxonomy also equals low accuracy we need a way to spell it out.

Success metric

One big part of knowing if the work we do even makes a dent is to define some success metrics before implementation.So how do we measure the success of the Activity view?

Company bet

Enriched transaction- and activity data can be viewed and edited/refined in a lightweight and understandable manner and we have proved that we can move work from internal to external users.

User value

Higher visibility for customers on their data as well as increased ownership over key editing functionality.

Measuring success

To validate our assumtion that customers want more ownership and control over their data one success metric is:
Number of external users performing released key edit functionality

To validate the company bet and prove it is possible to shift work from internal users to external we wanted to track:
Time spent for Data Service Analysts


Once we fine-tuned the design to sort out some of the UX-problems highlighted during testing, the Product Manager, the lead engineer and myself deviced a plan to implement the design in 3 different iterations focusing on key functionality.

The Biggest hurdle: Cross-team collaboration

Since our team did not own the infrastructure to join transactional and non-transactional data we had to rely on another team to build a new service for us (basically rebuilding Normatives data model from scratch). So to unblock my team, me and the Product manager had multiple sessions with this team to go through our needs as a customer of theirs as well as the needs of our clients. We still had a month of stand still where my team had delivered on their part but had to wait until the infrastructure was built to continue.


We rolled the activity view out to iteratively in percentages to our client base by replacing the previous "Supplier view" and so far we have gotten great feedback on the view on the points of using it to get valuable insights outside of our dashboards as well as how it empowers users to make changes themselves.

So how did we do?

1950 🎉

Edit-actions performed by external users (from launch + 90 days)


Unique external users performing actions

30% 👎

Increase in time spent per customer for Data Analysts

Actions performed

Still the majority of actions performed by internal users but we are seeing an upwards trend of actions performed by customers and the bulk of recategorizations are now performed by external users which has significantly reduced time spent per customer for our Climate strategy advisors. However the time spend per customer for Data analysts is tracking upwards but this is due to bigger clients with more complex data sets.

User traction

User traction of external users trended upwards and saw its peak in january (the blue line) now on a steady level of around 15-20% unique sessions per day.

Future vision

Next up is solving the more complicated edits and bringing a “co-pilot” to highlight what you need to act upon.
Thanks for reading!

← Back homeDesign system →