Media vs. Video Games Why is most media so bad and yet video games are so good? I mean that as a serious question. After all, video game companies are media companies, very much in the business of making money. And yet they have managed to completely avoid the hostility and avoidance that greets every other media company’s attempts at monetization. I believe the short answer to that question is that video game manufacturers are not weighed down by the legacy business models that are slowly destroying the larger media landscape. Therefore, they can adapt and (yes) market to today’s consumers in ways they find acceptable, engaging, and even respectful. Chris Dixon, a partner at venture capital powerhouse a16z, has broken down the lessons that the video game industry hold for other companies in the link below. I’m not sure that these lessons are really applicable to non-gaming companies without further disruption. But it’s worth a read. Application to Marketing: Video game companies use true freemium models to encourage engagement and loyalty. That’s free to win, not just free to play. Video game companies use crowdsourcing to create consumer interest and investment in their new releases. Video game companies release beta product in order to get valuable feedback from their most committed users. Video game companies encourage users to modify their games, rather than suing them for copyright infringement. In short, video game companies concentrate on providing value and entertainment and then monetize from their happiest users in the most transparent way possible. Think about the level of disruption that this kind of approach would create for every other type of media. Next steps: As disruption continues to plague the media landscape, it’s fair to ask if advertising-supported content isn’t the root of the problem. Read More Technical Debt and Advertising Advertising agencies and their clients struggle with technology. Not because there is any lack of intelligence or ability in the industry, but because the structure, processes, and approach of advertising agencies were created during a different time. When most agency executives were learning how to do the business, hierarchy and silos of expertise actually made a great deal of sense. Some visionary agencies have begun to apply the structure and processes of DevOps (Development and Operations) from the technology industry. But the advertising industry is still suffering from a kind of technical debt. Technical debt is a concept usually applied to technology start-ups that rush a minimum viable product to market, knowing that some of the technological short-cuts they took on the way will eventually need to be fixed for the business to scale. Application to Marketing: Thinking of the anachronistic aspects of the advertising industry as technical debt, actually provides a path forward for agencies. When a technology company suffers from technical debt, the solution is to bring in an experienced DevOps leader who is empowered to reorganize the staff and the processes, clear out old fiefdoms, and introduce best practices and modern software tools. The product roadmap will be delayed while the technical debt is being cleared up, but a company that refactors their code is understood to be building for the future. An advertising agency could do the same thing. But the biggest obstacle to change will always be the politics of an institution. Next Steps: There’s no harm in familiarizing yourself with DevOps. Read More Shopping on Pinterest Most apps don’t really know what a watch is. If you are using Amazon’s mobile app, you can take a picture of a watch and it will show you if it finds an exact match. But it has no concept of a generic watch. This is the problem of classification. It’s simply matching pixels, not matching the class of objects. Pinterest has been trying to drive purchase through their platform. Now they’ve introduced visual searching that uses deep learning technology, so you can take a picture of an object and they will try to show you similar objects that you can purchase. Take a picture of a watch, and Pinterest is capable of identifying it generically under the classification "watch" and then show you similar watches. Exact matches are unnecessary. One of the challenges of many “shazam for products” apps is that the picture you take needs to be substantially similar to the picture they store in their databases. If not, you get no match. Pinterest seems to be on their way to building something better. Application to Marketing: Pinterest is trying to become a discovery engine for products. Digital isn’t always good at discovery because it relies too heavily on recommendation engines which don’t reflect the serendipity of how people discover things they like in the real world. If Pinterest is successful in building and maintaining this deep learning platform, users will be able to find items they like out in the world and then purchase them through Pinterest. This would tie Pinterest’s monetization strategy to an actual product benefit for its users. That’s different from the advertising-supported model of most social networks. Next Steps: It’s worth making sure your products or your clients’ products are purchasable through Pinterest. Read More Learning to Decide In a recent opinion piece in Recode (link below), Technology investor Adit Singh suggests that self-learning software is the next big trend in enterprise software solutions. He is entirely convincing when he sums up the benefits of self-learning (or, if you prefer, deep learning) technology. I believe him when he says it would improve the efficiency and accuracy of decision making. I also don’t think it’s likely to happen. In technology, you get used to eliminating inefficiency. It is the culture of the industry. But big companies in established industries are structured to make the exact kind of decisions he hopes to eliminate. Automating a decision that could get you fired if it goes wrong is contrary to human nature. Self-learning enterprise software companies should concentrate on providing recommendations, not decisions. Application to Marketing: Marketing is one of those enterprise functions that places a high value on individual decision-making. It’s also a function that is currently drowning in poorly or incompletely analyzed data. Deep learning seems like a great match for marketing. But there are so many ways to get marketing wrong and the turnover for marketers is so high, I’m not sure that anyone will have the patience to allow the systems to actually learn, make mistakes, and improve. Next Steps: Deep learning is one possible solution to the crisis of big data in marketing. Stay tuned for developments. Read More