Behind our climate dashboard

Silja Mustaparta-Dolve
Oda Product & Tech
Published in
7 min readFeb 18, 2021

--

This is the origin story to Norway’s first climate dashboard for groceries. A tiny touchpoint with a surprisingly big impact.

As a company we care A LOT about data. Without good data, how would we make good decisions? We wanted to apply that same principle when communicating climate emissions to our customers: no greenwashing or judgement, but provide the relevant facts so people can make their own informed decisions. So, in November we launched a page where our customers can see how much CO2 emissions their orders produce.

What we’ve launched looks quite simple, but like always, making something seem simple takes a lot of work. It took multiple iterations for the dashboard to reach its current state, finding the balance between what is understandable, and what is correct.

Current version: The bars represent your shopping trips, and the colours refer to “very high emission”, “high”, “medium” and “low”. When you expand the bars, you see the emission details from each shopping trip, like in this example where a single pack of minced lamb meat causes 21% of this shopping trip’s emissions.

Where it all began: reaching climate goals without data

In Kolonial.no we have a goal of reducing our own carbon emissions by 50% by 2025. However, switching from plastic bags to paper boxes and shifting our delivery network to electric vans, only covers our own systems, while not addressing potentially the biggest source of carbon emissions: the food we sell.

And here our care for data kicks in. We lacked data on CO2 emissions from food purchases in our store. If we didn’t know how high the CO2 emissions were, or which products have the highest CO2 impact — how could we help our customers reduce their emissions?

Instead of “just” getting the data, it’s always easier in Kolonial.no to prioritize something if it has value for customers. The idea that got this all started was: what if we made a dashboard showing insights such as “time saved”, “trips to the store saved” and “climate footprint from food ordered”? That would force us to get the data!

The initial “fat marker” sketch from our Tech Lead Herman for the Climate dashboard pitch

The data reality, or the lack-of-data reality

There was however a reason we did not have this data in Kolonial.no already. There is no agreed upon standard in Norway for CO2 emissions for food, and no collected data. The Swedes have it, but does it apply well enough for Norwegian groceries? Are we even allowed to use it? We have almost 7000 products in our store, it could be a huge challenge to get trustworthy, comparable data for all products.

Chance would have it, Bob Van Oort — researcher at CICERO (Centre for International Climate and Environmental Research), had just contacted us on a possible research project. We took the opportunity to ask him — an expert on food CO2 emissions — to come and “tell our Summer students a little about emissions.”

Bob explained when it comes to CO2 emissions from food, scientists have already defined four groups: Low, medium, high and very high. So climate effect does not really come from choosing one apple over the other, it comes from choosing low emissions categories over high.

This was our key insight: It matters much more to compare across product categories, than products within the same category.

In addition: emissions from production has much greater impact than emissions from transport.

Bob’s graph: The disappointment of finding out cheese is in the high emissions category

Data that are accurate enough, and live!

This key insight allowed us to bypass the “we have no national data standard” problem by placing our products in four food categories, and giving everything within the group the same CO2-e level. Emissions would not be detailed on a product level, but as we had learnt the big win is not by comparing within the same category, but across categories.

The actual development job had been given to our Summer interns (hi Asgeir, Bendik and Sigrid!), and as Kolonial.no Summer interns have to ship something before they leave (no fake projects!), they worked like mad and launched an internal beta in August. Finally we had real data from real accounts! Suddenly concepts about dashboards turned into people’s real climate emissions, which got us all a lot more invested and feedback pouring in.

First version of the page!

How do we make the data understandable?

The main challenge during the Summer was how to display total emissions in a way people understood.

What does 56 kg CO2-e mean? Is that good or bad?

The first version solved this by showing emissions from the last 30 days of shopping, and comparing it to the average Kolonial.no customer, along with the Norwegian government’s target. If you don’t know what 56 kg CO2-e means, at least you can compare yourself to the average, and what the recommended level would be.

But, this had challenges:

  • Showing data for the last 30 days makes sense when you shop with us several times per month. But what if you don’t? We wanted to give our customers valuable data, no matter how frequently they shop with us.
  • Not every household is the same. When we compare customer accounts to an average Kolonial.no customer, big households will probably have a higher emission than average, and single households will have a lower. Even if the big household has a low emission per person, they might come off as “worse than average”.
  • Telling customers they’re “worse than average” felt a bit too judgmental. We wanted to go back to our original goal of providing facts, and letting customers make their own decisions.

Nothing compares to you, and Bob’s graph

The first version had given us a great starting point with real data, now it was time for our sustainability team to take the baton.

Since comparing yourself to others was problematic, how about making it possible to compare yourself to …yourself? If we showed emissions for each order, you could see your own trend over time, and easily impact the trend by shopping differently next time. Also, this way we have data for all customers, no matter how frequently they shop with us.

We still struggled with communicating the key insight, and ended nearly every user test having to show Bob’s graph about product emission categories. The reaction was always the same: “I knew there was a difference between different types of food, but I had no idea it was this big!”

We had tried to improve Bob’s graph, but user tests showed it was already pretty perfect. So we made a note on stop fixing things that work, and made Bob’s graph part of the page.

Bob’s graph with its new Kolonial.no look

How did zucchini get categorized as meat?

Data quality for 7000 products was still a huge task, verifying all products were assigned to the right category. Zucchini was first auto categorized as meat, since it’s something we also suggest for barbeques. That was fortunately an easy fix.

Some tougher calls were escalated up to our Climate Scientist Bob. Salami — is the emission level most like pork, or most like beef?” Beef, according to Bob’s data. How do we categorize mixed products, like frozen pizza or stew? Too difficult right now, we need to make a new “uncategorized” category.

Finally we were comfortable sharing the page with our customers. We “stealth” launched the dashboard, by putting it in Settings and then not telling anyone. We then sent a newsletter to some hundred customers telling them about it and a tiny questionnaire. 40 responses later, we knew the dashboard was understandable, but also that it was not obvious that each order in the dashboard was clickable. Quick design fix and we were ready for Christmas, and then for a proper announcement.

The small ball has started rolling

We are officially live now, and even though we are by no means done, we actually have what we originally set out to get: data, and thus a foundation to make our own informed decisions.

In addition to data, we also have a customer facing page informing about emissions, which is a more powerful asset than you might think. We lived well without it for years, but now that it exists, it is suddenly very hard to remove. Customers have started caring about it, they suggest ways to improve it and have started requesting making climate data even more visible. Now that the ball is rolling, it forces us to roll with it.

This fairly hidden page, has also taken dialogues with our suppliers to a new level. We were surprised to see how many are already working on getting emission data for their products, and who just want our guidelines on methodology.

Which leads to our main take away: how powerful it can be to get a small ball rolling. Doing climate related efforts is often a huge task, and it can be hard to get started at all. Doing something small, but visible, was our small ball.
— What is yours?

--

--

Silja Mustaparta-Dolve
Oda Product & Tech

UX Design manager at Oda, leading the Customer Insight Team