As someone who doesn’t enjoy Christmas (that’s a story for another day), I prefer to focus on spring cleaning instead, and in starting that process I’ve discovered this draft blog post. Originally written in May, I had just made the tough decision to resign and walk away from TreatOut, which also meant walking away from my two cofounders. I was soon to discover this included our longstanding friendship also. The months that followed have been pretty awful, the acrimonious fallout spewing onwards well into November. I‘ve learnt many lessons on how not to start a startup. I’ve only just found the strength to pick up again and start over with myFeastOut.menu. Boy have I got some material to write over the coming months.
But finding this blog post, in particular, has added poignancy. It was with a heavy heart that I read about the closure of Skills Matter at the start of November 2019. I am very privileged to have had the opportunity to work at Skills Matter, a wonderfully inclusive, safe and vibrant tech space where I learnt from the best of the best, dipping in and out of conferences, courses and events throughout the year.
It was where I was encouraged to write and get involved in public speaking (thank you Jenny), and it was a place where work colleagues also became friends. Skills Matter practised what it preached and was the first employer who dared to take a chance and give me my first commercial role just when I was about to give up hope. It launched my new career as a fullstack developer for which I will be eternally grateful to Wendy and Nick.
So while I continue to spend the afternoon clearing through all my notes and draft blogs I think hitting publish on this one is long overdue…
This month I took the opportunity to co-emcee with Jenny Martin, give a talk and host a roundtable discussion at Skills Matter’s newest conference, BeyondTech. It was a fantastic crowd and a privilege to be part of the discussion around the peripherals of tech. Here is that talk.
It’s Just Data
I’ve spent the past three years working on a side hustle, a tech startup I co-founded called TreatOut. I’m pushing to get full menu transparency from the catering industry. I’ve been told I’ll fail. Because it’s too hard a problem to solve. Because it’s too big a problem to solve. They tell me I need to simplify, to forget about all the ingredients and focus on the allergens. Keep it simple. I don’t agree. I challenge and push back hard every time I hear those objections.
When I first wrote this talk back in December 2018, I focussed the external pressures that the tech eco-system places on early-stage co-founders like myself. It’s common practice to simplify a data solution when you first develop an idea. We’re handed the Lean startup as our bible and I get this whole lean, agile approach to development, I do. But in our race to get a product out to market we often forget about the importance of the data we are serving.
In January 2019 Panorama aired ‘Takeaway secrets exposed’. Did you watch it? It’s still available on iPlayer until mid January 2020 and I urge you to take a look. It was a perfect example of what happens when no-one takes responsibility for the data provided. I watched that episode with mixed emotion and I’ve then re-worked this talk. I felt driven to defend my position as a potential disruptor and entrepreneurial technologist.
This talk isn’t just the theory about my own experience. Inspired by this Panorama episode, I’ll also discuss what’s happening right now in this particular vertical. Although it had me shouting at the screen in certain parts it fascinated me too because it highlights the consequences of simplifying and brokering data without responsibility. What’s actually played out over the past six months has really made me question what my true responsibilities are as a tech entrepreneur. I’ll start by giving you some context with my working example TreatOut.
What’s the problem?
I’ve picked this particular problem because it is my problem. And I’m pretty certain its a problem most of you will have come across. How many of you reading this now are, or know someone who is avoiding at least one ingredient? It could be because of a food allergy, a food intolerance, or it could be on medical grounds. It may be on religious grounds or even a lifestyle choice. The point is the reason itself doesn’t matter. But the sheer volume of people affected by the outcome is growing and growing fast. The reason is insignificant but the number of people with a need to avoid an ingredient is not.
And yet the current landscape remains pretty grim.
People with specific dietary needs tend to be treated as special. Friends call you ‘fussy’. Explaining what you can’t have to a stranger in front of an audience just to find out what you might be able to eat can get pretty awkward. People always start to ask questions about real personal information that you might not want to go into right there and then. It’s embarrassing.
But I just want to be normal.
Imagine not being singled out. If you could access the menu data you need in advance just turn up and sit down at the table like everyone else, without a fuss. After all, we celebrate with food, we comfort with food, we socialise over food. It’s central to our culture, to our very existence. We can’t not eat. If you stop eating, you die.
I just want to know what’s on the end of my fork, is that really such a big deal to ask for? Because this is just data, right? I sometimes wonder on days when in practice it feels like I’ve asked to move a mountain.
I’m having to sell the vision of what a future state might look like with TreatOut firmly established in the market place. As a founder you need to engage others to mobilise, get behind your product and forge key partnerships within the tech eco-system. I’m talking about fellow startup founders who are perhaps on the same journey. Serial entrepreneurs, angel investors and mentors from more established companies further ahead in their journey. Startup accelerators and incubators. This tech eco-system provides a vital network of support for early stage startups. And this is where the external pressures begin.
I need to test the market, to prove traction to investors. Other developers, potential investors, in fact by anyone who wants to have a discussion about TreatOut tell me to simplify the problem. “Don’t try to fix it all at once. Just focus on allergens to get something out there quickly.” But why pick allergens?
Because that information is readily available. Because it’s what we know, it’s what’s regulated, it’s what’s visible. But this isn’t just any old data we’re picking.
Sometimes complex problems do actually need complex solutions right from the start. As startups, it is often the case that we’re encouraged to keep it simple. But simple at what cost? What complexities are we removing? Maybe that complexity is relevant, maybe that complexity is necessary to actually solve a problem in a way that is useful to our users.
How do we actually determine the rules of the software we build, when people could be allergic to any ingredient? Currently, you can only have issues with 14 ingredients because those are the most common allergenic ingredients that are regulated. If you have any issues with anything else, this data isn’t going to help. How ridiculous does that sound? How on earth do our bodies know what is and what isn’t regulated?
We get so wrapped up in building the tech that we can forget we are ultimately building solutions for people. The body is a finely tuned, highly complex living structure.
Take the MacTwins for example, a perfect example of the complexity we are talking about. Tim Spector worked with them as part of a gut health research carried out at Kings College, London. They’re identical twins with 100% same DNA and yet their guts share just 40% of the same flora. Two identical people, two completely different guts with completely different needs. This is a biological complexity, not a complexity that’s borne out from developers overthinking a problem. This isn’t a complexity a tech team can magic away with brainstorming session and a pocket full of post-it notes. This is a complexity that is fundamental to our understanding of what we are actually dealing with.
We’re still trying to understand the connections between gut and brain function and the role that our gut plays in our general health and wellbeing. This will, I’m sure, touch all parts of the healthcare system, opening up all sorts of possibilities of using food as medicine and taking a more holistic approach to managing our own health on a day to day basis. Since managing my own health with food on a modified lowFODMAP diet (a lifestyle change not to be taken lightly nor without consultation from a registered dietician), I already think of food as my Superpower.
We are what we eat and yet we are not fully informed about what it is we are eating. It’s personalisation that’s desperately needed but we can only deliver personalisation if and when we start using the right data.
Media coverage over the past year around the tragic deaths of young people with food allergies has made us look at working practices and the reporting of allergen data within the Food Service industry and rightly so. In fact the government held a public consultation in the first quarter of 2019 and changed allergen regulation laws as a result.
But allergen regulation isn’t the real problem is it? The real problem here is that people are individual, we’re all different. The problem is that someone could have an issue with any ingredient, regulated, or not. So for me, this is not about allergens. That’s already been done many times over. There are many hidden complexities to consider when tackling this data problem. Stuff we don’t talk about as much. Stuff we ignore at our own peril. Yet it’s what everyone else in this vertical does. There’s plenty of restaurant apps out there based on allergen tagging. But they’re just not cutting it. It’s not enough, the regulated allergens are just the tip of the problem in ingredients reporting.
Creating the illusion of success
What is this magic bullet I’m told to use so I can demonstrate market traction? It’s data-scraping menus with their allergen information.
Data scraping is a method by which you automate processes to collect masses of data from across the web. It’s a popular method used by young startups to fill their own apps with data to create the illusion of having significant numbers of take-up, the traction that will attract customers, and the traction that investors always want to see.
But data scraping raises more questions than it answers. Where is the integrity in the data? What assurance is there that this data is relevant? When was put it out there? What validation is there? How do I even know I’m putting the right data into my app? Where does ownership of that data lie? How on earth can we expect end-consumers to trust that data? Because this is still just data?
What is just data?
Based on or behaving according to what is morally right and fair.
Just think about that for a moment. We are building tech that people use to make informed choices on a daily basis. As technologists, surely, we have a moral obligation to make sure that we are building technology using not only the right data, but using JUST data.
Look at what’s happened in the foodservice sector. In reality, many of the establishments listed on restaurant apps don’t match the information being presented to the end consumer. This most certainly raises questions on tech ethics. Restaurant discovery apps act as a silent go-between, positioned slap in the middle between the end consumer and the restaurant. Is it ok for an app to push forward information about these establishments, information which people are using to make choices, but not accept any responsibility at all for that data? Is the argument of merely passing data forward justified defence for taking that stance?
Challenging the business model
I’m still trying to figure this out and this is where it gets tricky. I think quite often the business models are a driving factor behind this questionable behaviour of ignoring responsibility. Spin it on its head for a moment, and instead of expecting a tidy commission from a business to claim or promote their listing, why not ask for a small customer subscription? What can be perceived as little effort still generates an income. Think about the implications in that shifting of finance.
It places the power in the hands of the end consumer. The onus is to keep the customer happy, building trust. In the case of the foodservice sector, removing the dependence on the restaurants as the main source of revenue allows for stricter controls on which outlets to partner with because that decision process will no longer be financially driven. It levels the playing field for smaller independents giving better support for local communities. An app is then people-driven, not corporate driven.
Nothing is without consequence
Moving onward to the issue of responsibility. Of course, it’s obvious that as a menu discovery app, TreatOut is not handling food but it will be passing on data. I’ve talked about this with industry regulators. They suggest TreatOut may be considered a food broker because it’s primary function is to provide information about multiple food outlets from which an end consumer can choose where to eat a meal.
Insurance companies don’t agree with that description label, but the discussion always points to the risk of a serious allergic reaction and the question of where legal responsibility will fall. In the case of a small local takeaway versus a successful tech app, insurers think the claimant would go after whomever has the most money in the bank. I know when the time comes for me to shop around for business insurance it’s going to be very hard to get.
Going back to the media coverage on the recent deaths of young people who died from unknowingly eating foods they were allergic to. There was a panic in restaurants and I followed the fallout across social media at the start of 2019 (#allergiesnotwelcome) caused by the appearance of the ‘may contain’ notices on counters. Kitchens handle allergens, we know this. We know that there can’t ever be a guarantee. We also know who the good guys are, who go above and beyond to make customers feel safe and welcome.
My concern was the growing fear on both sides from the diners and foodservice providers alike and the mistrust of tech apps, which in this particular vertical escalated at a worrying pace. What if we reach the tipping point and communication stops dead? We turn away people with specific dietary needs?
This sounds extreme but it’s real. A few restaurants have already turned away families. People are signing disclaimers. This is the consequence of not communicating the right data.
The issue of responsibility. I’ve had to retrain to work on this project and spent the past 2 years working as a software engineer in order to gain the knowledge I need to build the technology myself. I spent the first 12 months immersed in the heart of the tech community at Skills Matter, listening to the conversations about tech ethics. I understood during that time just what responsibilities I have in the passing on of that data.
I’m sure you are all aware of the swell of negativity and falling confidence in the tech sector. We’ve seen huge fines issued over anti-trust and data privacy breaches. Social media platforms are under fire on their responsibility to protect users against accessing damaging content. And of course, we can’t avoid mentioning fake news and political interference. In this landscape of mistrust, it’s no surprise that this is how Panorama tells us the tech disruptor is viewed.
“Tech disruptors move into an existing market, kill off the competition and then take over that market… with an attitude of asking for forgiveness rather than permission.”-Professor Andre Spicer, Centre for Responsible Enterprise
As a founder of a tech startup, this is where I leap into defence mode. As I see it, the aim is to grow a successful, scalable business. Building something from nothing takes hard graft and determination. There will be plenty of obstacles in the way that would kill off many a business, but entrepreneurship is about finding creative ways around those obstacles.
A new idea needs tough conversations and creative thinking. As a founder, you have to get right out your comfort zone to push boundaries and challenge established behaviour that as a society we’ve all come to accept, regardless of whether that behaviour is good or bad. If you’re first to market there will be plenty of others close behind eager to copy your success. If you don’t act fast to clinch the lion share of that market you won’t survive for very long. It’s a gamble and you have to have nerves of steel.
To get that market share you have to build the brand and raise awareness of your product. You need to spread the word, generate interest and grow your community on and off-line. The reality of building a tech-based product is expensive. You need to keep development in-house so that you can respond quickly to consumer demand for new features and bug fix any issues. A good team of developers is costly. It takes investment, serious investment. To secure investment you have to show traction. You’ve got to prove that you’re building something that people actually want, a product that they will use and pay for. Raising that investment is hard, and venture capital funds expect high returns.
Working in the startup scene is cut-throat, yes. But that doesn’t mean we can’t care. That’s why I want to do things differently, to get this right. To be responsible about what data I present to you. And it doesn’t matter if we are big or small. When we fill our apps with data, it can’t be any old data, it has to be the right data, even if that’s the harder route to take. And not only the right data but just data.
We need to forge partnerships and make sure that someone is taking ownership of their own data. We must make those people undertake to become guardians of their data, maintain their data, keep it up to date and relevant. Then, and only then can we expect people to have any trust in our data.
There are no shortcuts here. No simplification. Just responsible data.