What the research actually says about why async communication keeps failing
The meeting paradox
The async revolution was supposed to kill meetings, but it did the opposite. Since 2020, weekly meetings have increased 153%, despite every company adopting Loom, Slack, Notion, and a dozen other async tools. The tools multiplied, the meetings stayed. Something structural is broken.
The cost isn't hypothetical
Poor communication costs between $3,640 and $37,440 per employee per year, according to Axios HQ's 2025 report.
29% of workers say miscommunication from async setups directly strains their relationships with colleagues, and 15% have considered switching teams because of it.
The "nobody watched your Loom" problem is real and documented
Atlassian's own community forum ran a post titled "Dear team, thanks for watching my Loom video (I know you didn't)."
The post identified the root cause precisely: a Loom video competes with urgent Slack messages, Jira tickets that are already late, "quick questions" that are never quick, and if your video is 7 minutes with a 90-second intro, you didn't lose attention, you donated it to the void.
The performance problem is psychological, not technical
Even a Loom employee admitted: every time she recorded a video message, she felt uncomfortable in front of the camera, distracted by the process, wondering about eye contact, whether it looked awkward, and often re-recording multiple times to get the right take (Atlassian).
This is the person whose job was to promote async video. The bar for "good enough to send" is invisible, and people feel it even if they can't name it.
The real failure mode is the effort gap, not the format
Remote workers don't need another complicated system; they need something that solves one problem exceptionally well. Reddit threads consistently celebrate simple solutions over feature-rich platforms.
The problem isn't that people don't want to communicate via video because that is definitely booming. It's that the cost of making async video not feel embarrassing is high enough that people abandon it and schedule a call instead.
What actually works
Async describes a very specific structure: set context, establish reasoning, describe desired outcome, request a binary response.
With enough context and the ability for the recipient to either fix the problem or pass it on, responses come in under 20 minutes for most requests.
The structure of the message matters far more than the medium. A well-structured video beats a poorly structured one every time.
Most people never get the structure right because they're too busy worrying about how they look.
That's not a skill gap but a design flaw in every async video tool that's come before.
We built Velo because we think those two jobs: communicating clearly and performing on camera, should never have been the same job.



Replies
Separating clear communication from the pressure of preforming on camera feels like a genuinely fresh take on why tools like Loom fall short.
Velo
@laylabell Totally, I remember how many times I had to create a video, and did not because of getting my speech right. I have sent too many long texts because of this.
I amCurious how Velo guides users into that “good structure” you mentioned is it built into the workflow or more of a habit shift?
Velo
I know, communicating clearly and performing on camera should never have been the same job, what do you think?
@ajaykumar1018You're absolutely right! Communicating clearly and performing on camera do require different skills. Clear communication is about being precise and concise, while performing on camera adds another layer, like managing expressions, body language, and energy. Both can be challenging in their own right, but when combined, it takes a lot of practice and focus to make both work seamlessly. What do you think, do you find one harder than the other?
This is spot on the real issue with async isn’t the tools, it’s the cognitive load + lack of structure, so people default back to meetings. If Velo actually removes the performance anxiety + enforces clear message structure, it could fix the real bottleneck instead of jjust adding another tool.