Sundog Bio News | May 2025
Good morning,
We launched Sundog on LinkedIn this month, and we’ve been so happy about the response - plenty of exciting conversations to get deeper into in the coming weeks. Even more exciting, though, has been the progress we have made - both on our technology, and on our product.
Here’s the latest:
Alex Mitchell | CEO Update
It’s been a busy month of talking to investors, off the back of our Onstage application. It was great to be a part of their top 100 companies to apply. Exciting to see the deep engagement with the problem we’re solving, and it led to some really rich conversations. I’ve also been setting up time with scientists, to share our demo and understand how they’d use what we’re building.
This month, we finally got to a pricing model that we were happy with and that works at scale. Getting pricing right is going to be critical to our success - but we also strongly believe that our pricing model should be simple and understandable. Any potential customer should be able to look at our pricing page and reason about how much they're going to pay. At the same time, we want to be able to scale; our best customers, the ones who want to put all their images in Sundog, should grow with us sustainably, rather than costing us money. Now, while our entry point pricing stays the same, we're offering different tiers of storage. We believe our customers are going to see as much value in going back to their old experiments as they are in collaborating on new ones; and our new model supports their using Sundog well into the future.
A conversation with an industry scientist uncovered a new use case for Sundog - customer service. Experiments go wrong all the time, and instrument makers and reagent companies have huge customer service teams devoted to troubleshooting their part of the experiment. However, they struggle to look at the images their customer has generated, relying on screenshares rather than investigating the image themselves. Sundog’s asynchronous collaboration can free both the scientist and customer service, letting them work together effectively to solve their problems. We’ve set up demos to customer service leaders in industry to understand how they’d use Sundog, and what they’d want for the future.
Tom Armitage | CTO Update
I’ve spent the month working on the first Sundog prototype.
What we’re prototyping is the end-to-end experience of using Sundog. That is to say:
- uploading a data file
- turning that into Images you can work with in Sundog
- performing simple interactions on that Image: adding markers/counters, discussing things on it with other scientists/people
- turning one or more of your Images into one or more Figures: an output file you can put in a paper or on a poster
We’re doing this to learn about the product. There are two main things we want to learn:
Firstly: we want to show it to our target audience, and gain feedback on how they might use it, how they respond to it, to question our own assumptions, and discover the unknown unknowns - the things we don’t know - about the product and the domain.
Secondly: we’re learning about the deep details of the materials we’re working with. Are our assumptions about our technological approach correct? What are the hard constraints of the technology and data - and what we can we elide with good interaction design? And more nebulously: what does it feel like to do this work? What can we anticipate about the longer-term experience of building Sundog, based on this spike?
This all feeds into the how: working code, running online. There’s a time and place for paper-style prototyping - mocked-up interactions, assumption that things are possible. This isn’t it. Enough of the problem domain is new or esoteric, that it’s important we prove we can do the things we’re saying are possible.
What that means is: our image pipeline is “real”, working with real assets and code, built out of multiple services, that turns input files into images we can work with in the browser with no loss of fidelity. Our image viewer is real-time and multiplayer, letting many see each other’s cursors, markers, and comments, all updating instantly.
Quick aside - if you’re not a biologist, you might not be aware of the scope of the things I’m calling ‘Images’. We’re dealing with the kind of files that imaging tools like microscopes produce: these often contain many images, taken during the same imaging session. And they can be big. My current primary sample image is about 4GB in size, contains around 140 individual images, which are up to 512 megapixels large (25,000 pixels wide). And that data is all important. Suffice to say, our “imaging stack” is not just about resizing a JPG.
At the same time: almost everything else is in a minimum viable state - as little as possible to get the job done. Everyone tells you “if you’re not embarassed about your prototype, you’ve done too much”. Some of the work of prototyping is ignoring the emotive response you’re having to your own code and design - that it’s scrappy, fragile, tangled, embarassing - and staying focused on why you’re working so fast, why you’re cutting corners when you can.
We need to make the right things real so we can learn the right things. Learning is the number one goal. A lot of this code is going to get thrown away, and I’m fine with that; the main work of technology is not to produce code, but to understand the problem you are trying to solve.
What we will have, by the end of the month, is a compelling end-to-end story, that we can evaluate with users (and show to investors). It shows both what we can do, but more importantly, shows the possibility of Sundog.
Next month, we’ll share how our core customer - biologists - see the possibility of Sundog. We’re looking forward to it already.