Whose facts? When we talk about “tech bubbles” we are typically talking about over-valued technology companies. But technology can also create a bubble within which people live. Recommendation algorithms create a feedback loop which naturally pushes towards the extreme. If I click on one article in Facebook that is critical of Hillary Clinton, I am likely to receive other, similar articles in my news feed. If I continue to click, the algorithm will be less and less likely to provide me with news that contradicts my perceived viewpoint. The algorithm dislikes the squishy ambiguity of mixed perspectives, so it discounts the middle in favor of the extremes. All the better to target us. While we may wish for our opinions to be continually reinforced by supporting data, that’s not really very good for a democracy. Say what you will about old (elitist, left-wing, dinosaur…) media companies like the New York Times. At least they feel the need to challenge their readers biases with differing viewpoints. I may not enjoy David Brooks or Ross Douthat but I consider it my duty as a citizen to read and allow myself to be challenged by their ideas. Application to Marketing: Segmentation is important for marketers. Ideally, we would understand our consumer base on an individual level, but segmentation is often the closest we can get. In the old days, segmentation was largely invisible to consumers. But technology allowed us to place segmentation at the center of the user’s experience. We know what the consumer likes and so we provide them with more of the same. This is a mistake. First, people enjoy variety. I have essentially abandoned my list of Amazon recommendations because, while I enjoy science fiction, I do not enjoy science fiction to the exclusion of all other things. But excessive segmentation is also a mistake because it accelerates media fragmentation. Even if you don’t care if that is bad for democracy, it is certainly bad for mass brands. The day may come when consumers don’t know how to spell Oscar Meyer. It’s best if we share our facts, even the silly ones. Next Steps: If you use a recommendation algorithm, introduce a little more randomness. Novelty sells. Read More Friction and its discontents As Google, Apple, and Amazon dive into the voice interface, we encounter a world where technology offers less friction to the user. In Ben Evans latest blog post (below), we outlines the way in which the voice command given to an Alexa, an Echo, or a Siri eliminates the need to open an interface, type in commands, navigate and purchase. In this sense, the technology is almost frictionless. What we lose in choice, we make up for in ease of use. What I believe some of these technology companies have forgotten is that there is an older generation of Americans that prefer high-friction technologies. My father (a defiant and vocal luddite) is still outraged that the telephone company deprived him of his beloved rotary phone. The purposefulness of laborious dialing made every phone call more of an event. The friction allowed him to perceive control and mastery of the technology. With apologies to my father, I admit this is an extreme example. But mastery of any interface allows us to feel that a technology is somehow “ours.” Friction creates familiarity. Application to Marketing: Brands were built for a high-friction shopping experience. When I needed to go to the store and stand in front of a wall of detergent, the familiarity of the brand kept me from being overwhelmed. In a frictionless environment, what is the role of a brand? Do I need to remind Alexa to buy Tide? Or do I just ask it to buy detergent? And after 20 years of ordering detergent (or perhaps allowing Alexa to automatically buy detergent for me whenever I was running out), do I even remember the name Tide? And my children who grow up in this frictionless world, what meaning does Tide have for them? There is a deeper challenge for brands in a low friction world. I’m not sure they will have any meaning beyond product attributes like scent. Lovemarks they ain’t. Next Steps: Brands are made for friction. You can maintain your brand in the highest friction areas of the sales funnel. But that might not be the actual purchase. Read More Told you so I know, that sounds petty. But I’m so rarely right about something. And in the case of Facebook’s numbers, I was right. Facebook just announced that some of their ad metrics may be off by as much as 55%. I have been suggesting for years that Facebook is misrepresenting ad metrics. I also think they misrepresent user metrics and have no control of their bot problem, but that’s a whole other issue. What is shocking about all this is that none of the major media buying conglomerates have been publicly pushing them for greater transparency. Application to Marketing: As I have previously suggested, do not take any numbers coming out of Facebook at face value. Every six to nine months a news story comes out revealing some aspect of their reported data to be questionable. This is not an accident. It’s clearly part of the culture. Caveat emptor. Next Steps: Don’t believe their hype. Read More Fun times on the Mirai Botnet Prior to the election, a series of massive DDoS attacks managed to cripple certain vital areas of the internet. Then on September 30th, someone released the source code for Mirai, the botnet malware behind the attacks, on a public hacking forum. As security researchers dug in, they noticed some interesting details in the code. First, it seems that the malware primarily infected CCTV cameras. This is hardly surprising since many people do not bother creating a password for their webcams. Once it had infected the camera or other IoT device, a remote command & control function allowed someone to launch an attack. Oddly, within Mirai’s code there are a set of IP Addresses that it is specifically designed to avoid, including those of the US Postal Service, General Electric and the US Department of Defense. In addition, Mirai was designed to destroy competing bits of malware. Meaning that it wants to be the only malware hosted on infected devices. Finally, it should be noted that despite the fact that most of Mirai appears to use English, there are a few places in the source code that feature Russian language strings. This may mean that Mirai was either invented in Russia or whoever invented it would like us to believe it was invented in Russia. Application to Marketing: DDoS attacks can be devastating to companies that depend on ecommerce (either their own or third party) in order to sell products. An attack launched on a peak shopping day could cause a publicly traded company to miss earnings for the year. The best way to protect against botnets is to encourage people to scan their IoT devices and to use the available password protection. Next Steps: Install a password on your baby cam. Read More