Apple Glass, SMB TikTok ROI, AI-Content 'Dilution'

Apple Glass, SMB TikTok ROI, AI-Content 'Dilution'

Apple XR Headset the New Google Glass?

Following CES, there are reports that Apple will announce its long-awaited mixed-reality headset at the WWDC conference this year. The company is reportedly neglecting other devices as it focuses on this rollout. Rumored to cost $3K, Apple's headset will handle VR and AR but could be a huge misfire. Notwithstanding consumer surveys showing VR/metaverse "curiosity," and aggressive revenue forecasts, we won't see mainstream adoption of XR goggles any time soon (outside gaming). These products will find applications in enterprise and B2B contexts. But they're years away from broad consumer adoption, if ever. Witness Magic Leap's recent pivot to the enterprise, which follows Microsoft's earlier Hololens shift (similar for Google Glass). In short, 2023 is not "the year of AR glasses" -- or the metaverse.

Source: Midjourney

Our take:

  • There are mainstream uses of AR today that work well, mostly in apps (e.g., app-based furniture AR, Snap filters, 360 product views).
  • But glasses must fundamentally be affordable (and the form factor must also not be awkward). Simply put, $3K is simply too expensive.
  • Google Glass was goofy and too expensive. Though more sophisticated, the Apple headset still misunderstands the market. It too will have to find a home in the enterprise or risk being a spectacular failure.

SMBs Say Positive TikTok ROI

Even as TikTok is facing headwinds in the US Congress and state legislatures, it's increasingly being embraced by marketers, including SMBs. (We devoted a podcast to the TikTok opportunity.) A new survey (n=126) from software reviews site Capterra argues that small restaurants and retailers on TikTok are seeing a positive ROI and plan to invest more in the platform this year. Just under 80% said they saw a positive ROI – within six months (a long time). And 53% said they would increase their TikTok spend by 10% or more in 2023. Among the smaller subset that post organic content and advertise, 72% found organic posts to be "extremely valuable" to their marketing performance; 55% said the same thing for ads. In-feed ads, which are basically indistinguishable from organic posts, were found to be the most effective paid format.  

Our take:

  • Big survey caveat: the sample was only 126 retailers and restaurants. Capterra didn't disclose anything more about these respondents.
  • Also, 47% of TikTok SMB-respondents said they were going to maintain or reduce their spend in 2023. There was no discussion of this finding.
  • TikTok may be both the most accessible and most challenging marketing channel for SMBs. And going viral will become much harder over time.  

AI-Content Dilution

Here's a provocative hypothetical: As AI-content tools become more widely adopted by a range of professions, more online content will directly or indirectly be a result of its use. How will this affect the future quality of AI answers and content? (ChatGPT was trained on internet content.) While ChatGPT is impressive and digital marketers are busy documenting all the potential use cases, it can get answers wrong. And because there are no sources displayed, you can't evaluate their accuracy unless you already know the answer. (Neeva has a really good approach to this in the way it displays AI answers with sources.) I asked ChatGPT, "Will the quality of ChatGPT answers decline if training data sets are reliant on content that is increasingly created by AI itself?" Its answer was impressively complex and multi-faceted, rather than defensive – basically "it depends." The machine did acknowledge, however, that the quality of ChatGPT's continued output was contingent on the quality of training data.  

Our take:

  • Garbage in-garbage out is a universal rule of AI algorithms. The quality of training data is critically important to the quality of the output.
  • There is thus some danger that AI content will degrade over time and "dilute" or "pollute" future AI content.
  • Google faces a version of this problem as more (B2B) content is created with AI. However, human editing may mitigate quality problems.

Recent Analysis

Short Takes

Listen to our latest podcast.

How can we make this better? Email us with suggestions and recommendations.