5 bad tech ideas that development organizations just keep trying

I’ve spent most of my life working in technology and much of it working with nonprofit organizations, trying to help them use tech to make people’s lives better. In that time, I’ve learned that the first ideas social good leaders come up with for applying tech for good are usually terrible — a trait they share with leaders of tech for-profit companies, whose first ideas aren’t that great either. However, while the tech sector is particularly good at winnowing out bad ideas early, the nonprofit sector still has not learned these lessons. As a result, just about every week, I find myself persuading a social good leader out of one or more bad ideas. And they’re usually the same ideas — ones that I’ve heard over and over. It’s not that these ideas are always wrong; it’s just they’re wrong at least 95% of the time. So I’m sharing these ideas here with you in the hope that your development organization will have the chance to avoid these ideas at an early stage — and work out what might be effective instead. Bad idea no. 1: The app nobody will download Though 85% of the global population owns a smartphone, getting people to download and regularly use apps is hard. There are well over 2 million apps in the Google Play Store, but only a few thousand have meaningful downloads in a given month. Successful commercial apps are few and far between. They need to run a challenging gauntlet. First, the prospective user needs to hear about the app. Second, they have to be motivated to download it and try it once. Third, the app needs to solve a real problem for the user or offer something they really enjoy. And lastly, they need something which brings users back regularly — think social media, entertainment, and gaming apps. Even when a commercial app is downloaded, over 90% don’t get used again after just one month! And yet, there is a steady stream of social good apps which fail because they cannot compete with search engines, attempts to “recreate Facebook, but better,” or because a specialized app is not handy when a user actually needs it. Social good apps that do succeed have a powerful reason to be an app. If you need an app to do your job, like so many community health workers, it will get used. If you have lousy internet service, offline-capable apps might help. Bad idea no. 2: The one true list A frequent idea that comes up in the nonprofit sector is building the so-called one true list, such as every agency working in a given area. Unfortunately, people don’t go to list sites anymore because search engines are more efficient. But even if a comprehensive list is created and people actually come to a website and use it, that’s just the beginning. If a one true list is not well-maintained, it rapidly decays and becomes useless and irrelevant. Keeping the list accurate and current is a huge and costly undertaking most nonprofits cannot afford. Some one true lists have succeeded in aggregating mandated data, e.g., GuideStar, or because they have a huge network building them, e.g., Wikipedia, but these are the exceptions. Bad idea no. 3: ‘The next big thing’ in tech New technology arrives on the scene all the time, promising revolutionary abilities. Some break through to widespread adoption, such as cloud computing. A few hover for decades before finally making it, such as artificial intelligence. Many more settle into narrow niches — or flop. Nonprofit leaders are told all the time they need to leap into the latest technology. Here’s an awesome hammer, let’s go find some nails! However, the odds are low that the latest technology is the best solution to the problems nonprofits actually face. Choosing a technology because it’s “hot” rather than because it is the best technology for the job is a recipe for failure. Over the last five years, blockchain was an overhyped fad. Blockchain has some potential uses but for most nonprofit organizations it’s not matched with their database needs. And this year, generative AI solutions such as ChatGPT are causing nonprofit CEOs to ask me if they can lay off half of their fundraisers. My answer: only if you are good at raising a lot less money! Instead, be happy if your team can crank out 10% more proposals next year. Social sector leaders should sit back and let venture capitalists and for-profit companies blow hundreds of millions of dollars figuring out how to make the next big thing useful. Bad idea no. 4: The cult of the custom The cult of the custom springs from the belief, often sincerely held, that this one nonprofit is such a unique snowflake that it needs a customized software solution. This rarely ends well. If your organization depends on a custom piece of software, you are the only customer for that solution. And a software product is never finished these days: there will always be things that must be done. You are dependent on an outside developer to upgrade your software and fix bugs and that developer may not be as interested in maintenance as they were in the big project. As the only group paying for these changes, and with a limited budget, that means your software is unlikely to work that well when it’s launched. And it will go downhill from there! Restaurants and golf courses don’t contract with software developers to write custom software for them. They would much rather use solutions used by thousands of customers, where the resource invested in development is far greater and there is a strong incentive to maintain these shared solutions. Why should small- and medium-sized nonprofits without internal tech expertise go it alone? Nonprofits should only consider significant software development if they have strong internal tech expertise and a strong need that regular tech offerings don’t meet. There are an increasing number of successful nonprofits writing their own software, but they are still rare. Bad idea no. 5: The AI system that makes the problem worse AI is an example of a new technology that looked like a fad in the beginning but after decades matured into real value. Its value is dependent on advances in other areas such as the ability for computers to handle humongous amounts of data, better algorithms, and far cheaper computing capabilities. However, it’s not easy for nonprofits to apply AI successfully. They usually lack the data needed to make AI work for their social mission. And, it’s very easy to apply AI in a way that makes things worse. I hate to report this, but computers are still dumb as bricks. However, they are really, really good at finding patterns in giant piles of data, especially patterns that are really there. Unfortunately, those patterns include human biases and prejudices, which computers have tended to reinforce unthinkingly. Just think about how patterns of past discrimination have been locked into parole evaluation apps, “predictive policing,” and hiring apps that have an abiding preference for white men. What to do instead 1. Do your research before making a decision. Few social good leaders are tech gurus. Knowing about past mistakes is one of the best ways of avoiding future ones: Just say no! 2. Take advice from someone who understands you. If there is a tech nonprofit dedicated to your area of social impact work, try to work with them since they likely share your mission outlook. Or invest in an internal tech leader who will take the time to help you make better decisions about tech — because you will need it. 3. Start with the people you’re helping. If you must do tech inside your own organization, place people at the center. Especially those being helped. Talk to them about what they actually need and understand what problem you’re trying to solve. Once you’ve had those conversations, you can build something they actually will use to good effect, not what you think they should need. Tech is at its best when it is making people smarter and more powerful.