Problem Solved: How adam&eveDDB and IAMS Pioneered Nose Identification Tech to Help Find Lost Dogs

The NOSEiD app was made to solve a genuine problem: that 10,000,000 pets get lost in the US every year and that just a fraction of those make it home. 

Microchipping is a powerful solution to dog loss but in the US, it is not compulsory and so many pet owners don’t get their dog chipped. All that’s needed to create a nose scan is a smart phone camera and the NOSEiD app, making it simple and cost-free for anyone to register and protect their dog. It’s a real example of how technology can be used to solve a problem.

Nick Hirst, executive strategy director, adam&eveDDB> Mars Petcare is committed to its ambition of ending pet homelessness, and dogs going missing is a big driver of that homelessness – every year, 10 million pets go missing in the US, and many never make it home. As part of the Mars Petcare portfolio of products and services, the IAMS brand is introducing the NOSEiD app to help further that larger Mars Petcare ambition.

 

Sara Chapman, director of digital acceleration, adam&eveDDB> We know that losing your dog is painful, and research illustrated the scale of the issue. It also revealed that the process of reuniting a lost dog with its owner is time-consuming (at a time when the dog owner is highly distraught, they need to do quite a lot, quickly), fragmented and more laced with chance than you might originally think. On one hand, it’s not compulsory to microchip your dog or to keep the details on the chip up to date so you have a technology that works but that not all owners are engaging with. 

Drew Spencer, experience director, adam&eveDDB> Our research showed us that losing your dog is not something that dog owners necessarily plan for or read up on so when it happens, they are devastated and at a complete loss for how to respond. So there was a clear need for a solution that gave that emotional support and guidance too.

Nick> In research we’d also found out that every dog's nose is unique. So it was when we paired our ambition of supporting owners who lose their dogs with the idea of using a dog’s unique nose print that we realised we were onto something. We started to conceptualise whether we could develop technology that would scan dogs’ noses, register them on a database, and allow anyone else with a smartphone to scan a missing dog’s nose and identify it. 

Sara> Once our initial tests showed it was possible, we knew that we were on to something. Then all we had to do is work out how to make it!

Sara> We were really excited by the thought that the nose print could become a biometric identity and that the nose print could be linked to details about the dog that would help someone to identify it if it got lost. 

So our first step was technical exploration; we needed to prove if it was possible.

Our development team spent hours analysing and categorising noses until we managed to prove that we could recognise and categorise different elements of a dog’s unique nose in a way that enabled us to tie it to its identity. Once we had that proof of concept we knew we had something we could scale.

Drew> Next, we developed wireframes to outline the basic user experience, which we tested with a group of dog owners and another group of people who had been involved in helping a lost dog get home, using mobile ethnography. This helped to inform the additional features and functionality the app would require to be useful and usable. 

A favourite quote from that moment was when one pet parent relayed to us about the time when his daughter spent nearly six hours drawing their dog, because they realised they didn’t have a recent photo.

Sara> Early on we all agreed that it was really important to us that whatever we created was easily accessible to every pet owner and consequently, needed to work on mobile phones. 

It was also important to us that what we put out there wasn’t just technology; we wanted to consider the wider user experience and ensure that what we created would support the pet owner through the highly emotional experience of losing their pet and help them to take every step they could to help bring their dog back home. 

Drew> We wanted the user experience of the NOSEiD app to be like that of a trusted neighbour or close friend – someone who loves your dog and, in your time of need, is there to help. Someone who would help with the logistics such as gathering important details, asking you questions you might not have thought about, keeping an eye out for reports of dogs that have been found nearby, and even designing a poster for you to stick on some telephone poles around the neighbourhood or share on social media.

We also knew that, although the nose scanning technology is impressive, that we would need to put that into the context of a wider experience that is about capturing the details about each dog that makes them unique. Each dog has a profile that is more like a profile on a social network than an ID card. You can get a sense of who they are and why they are special to their owner. With microchipping, it’s easy to see the dog as a number on a sheet of paper (that is sitting somewhere in a file cabinet drawer that you haven’t opened recently). With the NOSEiD app, a dog is a unique pet, just a couple of taps away on your phone, who is loved and missed terribly when they go missing. The process reinforced this with us, as the dog owners we tested this with would get emotional when just being asked to imagine their dog going missing.

Nick> The early period of this project was very focused on experience and technological “ideation”. We knew what we wanted the NOSEiD app to feel like and how amazing it would be if a simple scan would match a missing dog to its owner.

Defining the way we wanted the experience to feel was a relatively straightforward process; our early user research had revealed some key user needs and Drew, our experience director, had a very clear idea of the experience we needed to bring to life in the product. Drew worked with the team to establish clear principles, user stories and a series of iterative prototypes that allowed us to get to what our MVP solution would be.

Drew> Working out the exact features to prioritise was more tricky. There were a multitude of ideas and we really had to be disciplined with ourselves so that this product didn’t get over complicated. We set out very early on that we wanted the app to be a utility that would deliver value to the pet owner and help them protect their lost dog. This focus enabled us to focus time on making an experience that delivered the core experience well.

Sara> We knew that asking a pet owner to scan their dog’s nose was a new behaviour that we’d need to guide them through. So, we looked at a lot of early stage technology and how they ‘taught’ us new behaviour. We also talked about our first experiences of using new technology e.g. the first time we took our fingerprint for TouchID on our iPhone, the first time we scanned in a credit card rather than typing in the number. You’ll see things like the feedback and guidance given during the creation of the nose scan takes from these design languages.

Knowing that we wanted to provide both practical and emotional support for a pet owner 

meant that we looked at a really wide range of inspiration to guide the overall experience. We looked at everything from products offering emotional support at a time of crisis to dating and friendship apps and how they present the moment you’re matched to give us inspiration.

Sara> The user testing process was also fascinating as it really made us challenge some of the assumptions we made during the early ideation and when we created the prototype; for example the digital community often talks about streamlining processes like registration to avoid asking people to fill in lengthy forms but what we heard from the beta community of dog owners we’d recruited is that they felt reassured by a product that asked for more detail rather than less. It gave them the sense that everything they’d want someone to know about their dog had been covered.

From a technical point of view, one of the key challenges we had to solve very quickly was to find the data we needed to train the machine learning component of the product; in order to scan and recognise a dog’s nose, we needed a big set of data around what they looked like. Initially, we compiled a database of images of dog noses from image banks to start training and used our own dogs to help us train the ML, but, if we were going to use this tech to save dogs, we knew that we needed to test it on more live dogs.

So we created a very bare bones prototype and recruited a “scan squad” - a group of people who could seek out dogs in parks, and dog hot spots across London and scan their noses to give the scanning tech and intelligent matching behind the NOSEiD app a wider amount of data to learn from.

Sara>  Working with the beta community brought us one of the most interesting challenges of the project - dogs themselves. When a human goes to have an X-ray or an MRI they know to stay still, but scanning a live animal is a very different prospect, and our beta community showed us that whilst some dogs were very chill and obedient, others would sit to be scanned for a lot less time.

Nick> From a technical point of view, we wanted as long as possible and as many scans as possible to create the most accurate nose scan, but from a user experience point of view, we needed creating the nose scan to be something any pet owner could not only accomplish but would not find too difficult.

Consequently, we needed to find the perfect balance between getting the most detailed scans of the nose we could and the optimal experience for the pet owner. Working with owners, we discovered the best scan duration that enabled us to be accurate and also for it to work with more energetic dogs. In the process owners also shared a lot of their tips and tricks with us, and we’ve incorporated a few of these into the interface – like the ‘squeak button’. This came about after some dog owners told us an anecdote about using a YouTube clip of a squeaky toy to get their dog’s attention. We thought this was genius and put a button to recreate this effect into the app itself.

Drew> The emotional response from people when they saw that we included features like an “Automatic Poster Generator” (creates a “lost dog poster” that can be printed or shared to social straight from their phone) was overwhelming. People who had actually lost a dog before commented on how many little things we’d done to make that part of life easier. It was great to hear comments about how we’d focussed on the needs of the owner, not just the dog during such a difficult time.

Nick> Mars Petcare has an amazing network of pet specialists who have helped us to shape the product. During development of the product we were able to speak to several of Mars’ shelter partners in Nashville who gave us amazing insight into the types of things that help them identify a lost dog when it comes in to them and the process that many pet owners who’ve lost their dog go through.

We’ve also partnered with Psycle Interactive, an app development specialist on the product development and partnered their developers with our CX, UX and design team to develop the solution.

Drew> We’re launching in Nashville as a beta test; although we’ve tested continuously throughout the project we’re now really excited to see the product being used by a wider range of users and to look at live data and user feedback to help us steer where we go next.

Along the way, we have tested with users across a variety of methodologies. From paper prototyping and video interviews over a mobile ethnography platform to larger stress-tests of the technology build. Testing and iteration has informed every stage since the very first brainstorm.

Sara> Live data is a huge factor determining where we go next. Whilst we have a really exciting roadmap planned, we’ll be reprioritising and adding to that based on what the data tells us. 

Nick> Now we’re live, we’re pairing up our experience specialists with Numbers, our team of analysts to identify what users are telling us about the product and to identify how we optimise and build out new features to respond.

Sara> There wasn’t a “standard” process or solution and we had to think creatively to find a path through and put a plan in place even when things aren’t 100% certain. For example, how do you work out how to test a proof of concept piece of tech without starting on a full app build, or how do you deal with the fact that there isn’t a standard way of scanning an animal’s nose yet?

People often talk about working like a start-up to get this type of work done, but really success lies not just about embracing lean or agile or any other of the things we hear about from startups, but in creating a team dynamic that embraces uncertainty and acknowledges we may not know if this is going to work but that we’re OK with that and we’re going to forge ahead anyway. 

Nick> Mars Petcare’s ambition is ending pet homelessness, and the NOSEiD app is a genuine solution to help solve that problem. The goal is to keep making it better to help ensure every lost dog finds its way home.