Cleaning Up Our Mess: Digital Governance for Designers
The Internet and Web have reached a tipping point. We’re now witnessing the surfacing of harmful patterns and norms that we designed—often unintentionally—into our products, services, and communities, and the world we live in. Designers who work in the enterprise are, like their peers in startups and big dotcoms, vulnerable and culpable and need to consider some big questions: How well do we manage our data? How inclusive are our development practices? How broadly and deeply do we think about the impact of what we build and deploy before we scale it for our customer base? We need to move forward with intent. We need to govern our digital spaces. A necessary first step towards that goal involves designers examining—with honesty and introspection—our role in the creation of what’s online. The World Wide Web is nothing more than the accumulation of what digital makers have put there. We made this mess, and we need to talk about how we are going to clean it up. Digital governance expert Lisa Welchman will reflect on how 25 years of passionate and agile web development got us where we are today, and the consequences of the lack of self-governance by the digital maker community. She will show us a path forward from this mess, outlining questions we can ask and steps we can take to govern better what we have created and what we will create in the future.
Lisa hails from Baltimore. What she is talking about today and the history of Baltimore are intertwined. At the end of the day, she is going to try to help us understand a broader sense of standards – beyond design standards, sensible standards that keep us safe and operating well online.
There was a Great Baltimore Fire of 1903, which happened in the early days of firefighting equipment. The community surrounding community wanted to help Baltimore, so people came from Philadelphia, New York, and Virginia. But the coupling from the firehoses weren’t standard, so there were limits to what the volunteers from other areas could do to help. So, you can have the right intentions, the same values. But if you don’t tune it properly, you don’t get what you want. On the Internet, there are plenty of well-intentioned people. She has worked in really large enterprises, start-ups, educational institutions, and government agencies. They have good intentions, but those don’t always manifest well – they don’t turn out the way people wanted them to.
So that is what she is going to talk about today:
- Digital Governance 101, so we know what she means by that
- The mess we made online, how we got that way
- What ought we to do when we leave here, as individuals and as a community
She is not a designer, but she does design people systems. Her background is in the enterprise; she started working coding web pages for Netscape when she was home on maternity leave. She then worked for Cisco systems in 1996, publishing product pages. Many of those pages are online in basically the same structure as they were back then, which shows you how fast things move in reality. Cisco was very forward-thinking in 1996 – they were dealing with multi-lingual, multi-country, etc. So she got a crash course in what could go wrong. Cisco understood digital, they got it – but they still couldn’t run their web stuff . IT and Marketing were fighting, there we arguments about who owns the home page, if different countries could do things differently. There was so much infighting, and it wasn’t that the CEO didn’t get it.
So she left, starting easing towards digital governance. She focused on tools first, then later she focused on workflows. Finally she realized it was the people – they were disorganized and didn’t know what they were doing!
Governance is ultimately about decision-making. It is not workflow processes, or who gets to approve what before it gets published. It’s a subset of corporate governance. Marketing and IT both think they own it, then stuff like metadata falls on the ground. This is an organizational concern, because o-one knows who is supposed to decide what around digital. You first have to determine who (it can be multiple people) are is working on this. It does need to have qualitative and quantitative aspects. It is usually not you – it requires very senior business people or you can’t operationalize your strategy. The policy provides guardrails, should that be IT? Brand is Marketing. What about things like GDPR? Policies keep you going, keep you on track. What needs to be written, who needs to be writing them? She occasionally gets called by a designer who has written rules, and wants her help in getting people to follow them.
Most people she works with don’t know these answers. But they need to be operationalized and made to stick.
We need to move from an informed strategy to clear policy, to sensible standards. It is not a by-product of a project, it is there all the time you need to keep iterating on it. That in turn requires targeted and intentional collaboration. Good governance prepares the field for targeted and intentional team collaboration.
You can visually understand what you are trying to execute, but now you need to architect a team to get it done. The core are the people in charge – writing strategy, policy, and standards. The distributed team is out there making stuff, oftentimes creating shadow IT, shadow Marketing, etc. and driving the core team crazy.
If you haven’t designed this, and you’re not sure who is working at each level, it’s not going to work. You can’t have people complying with policies and standards if you don’t know who they are. Oftentimes, external vendors are doing more than the people inside, so you need to set up appropriate communications with them as well. They set up tons of microsites – it’s a business model but it’s not in your best interests.
The right framework can drive any type of operational scheme (slow vs fast, loose vs tight), but they need to be designed with intent.
Talking about the mess
There are three potential areas of mess:
- Enterprise properties and channels. You can look at what a company has online and tell a lot about them. If they don’t have Single Sign-on (SSO), that’s an IT issue. Another client had a site that looked good, but when navigating to the second level, was no way to get back to the home page. This was the web, but it’s the same thing for social, anything digital. No standardization results in people fighting, and it makes the work crazy.
- Internet and the WWW. Privacy and data concerns are a big focus at the moment. But in reality, some of these dot-coms are providing air cover for us. Enterprises are not much better, but it just doesn’t manifest as visibly.
- Algorithmic mess. There are many aspects to this. One small example is the fact that facial recognition for border entry doesn’t work well for dark-skinned women.
Enterprises do have areas of governance, but they lag in getting things online. Dot-coms are generally the opposite – they put lots of crap on line but they are organizationally immature.
Technology gets rolled out, things happen, it gets messy, and it needs to be fixed. Television was that way, radio, early computing. Each moved through a maturity cycle. She is going to talk specifically about how that happens.
Somebody creates something new, and others try to mimic and improve on it. But they are all trying to come up with MVP. We think that 25 years is a long time, but technologies take 100-150 years to get to the point where they are done. But today we act like we’re done, but we’re still trying to get to something that works, like a set of web properties. One of the challenges is that the technologies that feed into those properties are also on their own Eureka cycle. Over time, we start to see issues – the multiplicity of nonsense. And then, we get into many debates and arguments.
We do always end up governing, so the real dialogue is in the how, agreeing on the right way to do it. Eventually we arrive at standards. The real (sad) truth is that when we get to proactive, safe development, everyone is bored. It’s repeatable, it scales well. Then it’s all about pricing. At the very beginning it’s a few very cool people, but in the end it’s deep – lots of people. Right now we’re in the painful period where things need to change.
She is going to walk through an example from the U.S. automotive industry. The first gasoline powered car was demoed in 1875. It was driving around, but it was not particularly safe. There is a lag between those innovations, the defining of standards, and eventual safety testing. But eventually legislation was developed. There is a substantial gap in time between the first gasoline powered car in 1857 and crash test dummies in 1971. The web is similarly complex.
So, how do we start the fix? Lisa shared an excerpt from an MIT lecture, which explained that you can’t just change the algorithm – you have to change the organization. He talks about a ‘valid system’, but valid doesn’t mean it’s right; just because it runs doesn’t mean it’s ethically or morally right. So, we are the fix. Content doesn’t just make itself. “I can’t believe the algorithms are doing this.” What do you mean? Of course they are doing this – someone made them that way. Everyone has biases, so this is not an accusation. It’s not about us as individuals. It is an intersectional problem. Figure out where you are going to play in the array of possible solutions. It is not happening to you; the reality is that we are making it happen to everyone else.
Things we can do
Lisa articulated practical things we can do to address the mess(es) we have in front of us:
- Define digital safety. What would it look like, how would you measure it? What would a crash-test dummy equivalent for digital safety look like? And when in development do we consider that? A lot of the reason this isn’t great is that designers don’t understand the data structure it’s operating over. So, we need a more diverse set of people in the room. Who will be engaged in those discussions?
- We need to re-articulate MVP. It works, it performs, it’s a good experience. But is it inclusive, moral, and safe as well? We can’t afford to be narrow in our definition of MVP any more.
- Pick your cause. Digital governance is about the enterprise, or focus on the web – be part of W3C councils, for example, or The Web Foundation. If you care about internet governance, you could join the Internet Society. When conversations about Net Neutrality were underway, we (designers) were not engaged. What does it mean to uphold human rights online? She has been working with the UN on that. There is a big gap in internet access between men and women in the developing world, for example. How can we address that?
- Think about ways to be generous. We are really important, which is why she is picking on us. Volvo developed a three-point seat belt, and they gave it away because it was better to save people’s lives. Think about being generous in that way – don’t always think about how you might monetize something. We all make our living off the internet, but it was given to us. The culture of digital should be inclusive and sharing – because that is ultimately at the foundation of what we do. Vint Cerf and Tim Burners Lee were speaking in Washington DC. Vint was like the funny granddad, the internet is now mature “every bad think that can happen in the world now happen on the Internet”. Lisa’s amendment to that is “And every good thing, too.” The capacity for good is so amazing, let’s participate in that, get involved.
People speak in a positive sense about new technologies. The idea of disruptive technologies. Why wouldn’t we want harmony and collaboration instead? Do new technologies always create chaos?
Disruption is natural, it’s an evolutionary process. The Honda Accord is very safe, it keeps it’s value – but it’s a commodity ripe for disruption. Eventually you reach that tipping point, and there will naturally be some messiness. Eric Fisher has something called ??? innovation. Where in the process do you consider ethics? You don’t want it so tight that people can’t event, but you have to consider safety before you scale. We don’t know what that point is with digital, because it is so new. There will inevitably casualties, but as we scale and mature, can we make them safer. That will eventually be disrupted again.
If change is a constant and we can take that as a standard, have you ever seen governance or legislation where it’s time-boxed? Could that be a strategy for approaching governance?
Product is at the organizational level. Your governance framework has to be at that level. It should encompass a variety of technologies, design standards. So, you can engineer something that has flexibility in it. It’s an organizational concern because you need to clarifying decision-making, and create an organizational scheme to support it. It might be super loose for websites, and tight for social. It’s an array of things going on. You have to make sure the team you’ve designed is appropriate. It’s not about tactical review, or choking out new ideas. If you hire well, then you shouldn’t be that controlling.
In the case of algorithm bias or other things, what is the role of testing and feedback?
This is not about the outward facing aspects. She knows that as designers we care about the outward facing part. This is about the inward-facing part. User testing can be helpful for awareness, but it is after you’ve built something to test. We are trying to manage disruption in the enterprise. It’s about getting the internal machine organized. We need to rethink how we reward employees – not reward along silos.