Remember Lenny

Writing online

You are here: Home / Archives for rememberlenny

rememberlenny

by

Scoring in-person audience engagement through video footage analysis

Imagining how to provide speakers and performers with insightful feedback

Originally posted on Medium here.

Speaking to a large audience

Speaking to a large group is stressful. Something you can leave thinking things went horribly, when you couldn’t be more wrong. If you’ve spoken in front of a large group before, then you know what if feels like to not know what your audience thinks. While delivering your message, you can become so focused on one person, that you misinterpret the group as a whole. The one person texting or falling asleep can influence how you read the audience.

You could completely misunderstand the audience.

Image what it would look like to be able to record and reflect on your audience, based on what you say.

Quantifying the audience

First: Take a video camera or cell phone and plop it facing your audience. Have the camera film your audience during your segment of speaking or performing.

Second: Take the recorded footage and drop it into some non-existent, but feasible, software that analyzes each audience member’s face and body over time. The software could designate each person a score for their emotional state and level of engagement. Based on the scoring, timelines of scores could be generated for each person and the audience as a whole.

Third: Summaries from the scores can be turned into recommendations or insights for the speaker/performers to take into consideration. Based on these recommendations, a feedback based improvement cycle can be developed for speakers/performers, and a quantifiable score could be attributed to each audience.

All of the software to implement this is readily available.

Review

Imagine knowing how many audience members stayed for the entire time you spoke? Imagine knowing when people were most engaged or most disengaged. Reviewing footage is one thing, but being able to compare your results to previous engagements would be priceless.

Who would care

If it existed as a service, large venues could offer the insights as a service. Comedians could use the results to help refine their craft. Corporate speakers and salesmen, who travel on behalf of their company, could provide quantifiable results based on their group sessions.

Thoughts, concerns, or comments?

This is an idea. I’d love to know what you think.

Originally posted on Medium here.

https://blog.rememberlenny.com/2017/01/14/scoring-in-person-audience-engagement-through/

Scoring in-person audience engagement through video footage analysis

by

Large groups are hard to read. By using software, the engagement of large groups can be quantified for future review.

Imagining how to provide speakers and performers with insightful feedback

Speaking to a large audience

Speaking to a large group is stressful. Something you can leave thinking things went horribly, when you couldn’t be more wrong. If you’ve spoken in front of a large group before, then you know what if feels like to not know what your audience thinks. While delivering your message, you can become so focused on one person, that you misinterpret the group as a whole. The one person texting or falling asleep can influence how you read the audience.

You could completely misunderstand the audience.

Image what it would look like to be able to record and reflect on your audience, based on what you say.

All is takes is a cell phone facing the audience.

Quantifying the audience

First: Take a video camera or cell phone and plop it facing your audience. Have the camera film your audience during your segment of speaking or performing.

Second: Take the recorded footage and drop it into some non-existent, but feasible, software that analyzes each audience member’s face and body over time. The software could designate each person a score for their emotional state and level of engagement. Based on the scoring, timelines of scores could be generated for each person and the audience as a whole.

A software dashboard could break out each member of the audience.

Third: Summaries from the scores can be turned into recommendations or insights for the speaker/performers to take into consideration. Based on these recommendations, a feedback based improvement cycle can be developed for speakers/performers, and a quantifiable score could be attributed to each audience.

All of the software to implement this is readily available.

Review

Imagine knowing how many audience members stayed for the entire time you spoke? Imagine knowing when people were most engaged or most disengaged. Reviewing footage is one thing, but being able to compare your results to previous engagements would be priceless.


Who would care

If it existed as a service, large venues could offer the insights as a service. Comedians could use the results to help refine their craft. Corporate speakers and salesmen, who travel on behalf of their company, could provide quantifiable results based on their group sessions.

Thoughts, concerns, or comments?

This is an idea. I’d love to know what you think.

Idea: Slack/Github activity monitoring

by

Idea: Slack/Github activity monitoring

Working for a distributed company has made me aware of an issue: employees can “disappear” too easily and no one notices. When employees have issues, there is no predetermined face-time to nip it in the bud. HR visits, employee gatherings, and good ol’ one-on-ones are easier to avoid. Distributed teams naturally focus attention where work is getting done and miss potential areas where employee satisfaction is not ideal.

Poor employee engagement could be pegged to a metric based on the interactions a person has per week.
I think you can solve for this without being creepy.
Its plausible that a basic observation of how users interact with one another in a public venue is a high signal for employee engagement. Using a Slack based company for example, if users are never active in any Slack channels, then they are most likely not participating in group discussions. This could be the result of many things, but for one, it could be a individuals sense of no longer feeling engaged by their coworker community.

Employee engagement viewed over time.
It is important to note that a person’s engagement could be interpreted many ways. A highly participatory user could be feel very disengaged. A non-participatory user could be very active in syncing up with channel activity, and therefore feel engaged. Based on these examples, public participation is not foolproof.
The public participation could be a strong metric when comparing a user against themselves, over time. For example, when a user begins at a company, their activity may be above normal, due to the high number of introductory engagements. As time goes on, an employee may shift into a participatory trend that resembles more alongside their overall behavior. Based on the corrected average, any significant veering away would be a event worth noting.

Commit history mapped against a user’s commit trend could surface potential employee-project connections.
The same analysis could be run on an engineering centric organization with a shared code repository. Whether Github or some locally Mercurial equivalent, the individual contributor trends could be interpreted overtime for employee-project match. For example, potential fit can be gauged based on an employees rate of contribution during certain projects.
This can be creepy from a management and HR standpoint.

User activity could signify user happiness.
To make this work, observed results should never be used for consequential results. Punishment or reward based on analysis would be counterproductive. Employees who believe their overall behavior is being monitored and tracked could feel untrusted. This form of analysis could be the cause for employees to actually feel dissatisfied.
The overall goal of this would be to catch the otherwise uncaught signals. The goal is not to monitor your employees. If something major changes, some structure should be in place to see that the a struggling employee doesn’t simply fade away due to no longer being in the lime light.

Computer Vision in commercial establishments

by

Security camera to scene analysis and summary insight report.Computer Vision in commercial establishments

Since the inception of the “quantified self” and corresponding hardware devices, I have been curious about what people are doing with the data they collect about themselves. Fitbits, Apple Watches, and step tracking tools provide individuals with insight about their physical activity. With the increase in data generated about individuals, I tend to wonder how this information translates to groups.

Hardware companies have written interesting studies analyzing the data their platforms and devices collect, but this information is restricted to the company’s chosen points of interest. Smaller establishments, such as restaurants or commercial stores, do not have the benefit of gaining insight from the increase in these data points.

I think there is an interesting way to think about the other sensors that are newly available for analysis, specifically, the sensors that have always been around, but have not been fully utilized.

Tracking trends in public spaces is now available to all

In years prior, sonar, lasers, and similar one-use sensor technologies have been required for computing the physical world with data. Tracking the traffic trends in public spaces or tracking the foot traffic in commercial establishments have previously required specialized hardware. The specialized hardware was available to cities and large commercial ventures, but again, not for smaller scale commercial ventures.

Security cameras in restaurants, cafes, and bars could be used to reconstruct scenes.Software advancements now make common use digital cameras into complex sensors. Using software that processes images, the content of an image can be quickly analyzed for recognizable objects and human behavior. Computer-vision based software can be used to reconstruct scenes using networks of cameras. This allows for common security camera infrastructure to be used for real-life google analytics like analysis.

Data gained from analyzing customers could be used to gauge customer loyalty.Human activity in a common space, monitored with a digital camera, can be used to interpret common movement into analyzable data points. Further, in commercial establishments, transaction data can extend these reference points to understand how a brick-and-mortar business operates. I assume this information would not provide revolutionary insights for a business owner, but would be a great reference over time for change in business activity. It’s reasonable to assume this data, over time, would help to foresee trends in business growth or decline.

In the immediate, this kind of monitoring and analysis is uncomfortable. I’m curious to see if our expectations and cultural norms will change with time. Already, our activity online is analyzed and used as a commercial product. The click-trails we leave behind are aggregated and used to better target ads. The search behavior, communication patterns, or social graphs we create offer greater insight into the kind of person we are than most of us would like to admit. This monitoring of behavior is something people have become both more sensitive to and less aware of.

Given the changing times, I wonder if brick-and-mortar businesses would be comfortable using their existing infrastructure to gain deeper insights in to their business. Further, I wonder how customers would feel knowing that their activity is being monitored and analyzed for the benefit of a business.

This technology is already being widely used in public places. Companies like Sidewalk Labs that offer free wifi using publicly placed kiosks outfitted with high quality cameras, which are undoubtedly being used to monitor public activity. I’m curious if this kind of monitoring and analysis would have required a different approach using older hardware solutions in the past.

Also posted on Medium

The history of art and the finiteness of human production

by

There is so much ART, but its actually limited. A computer could understand all of the art humanity has produced and analyze it down to its features. Color, stroke, contrast, characters, symbolism. All the symbolism that is produced for meaning can be interpreted.

Computers could understand the components that underline meaning in art. Computers have all of the historical context and relational knowledge to make connections people dont make. The internet’s resources provide all of the resources to interpret cultural value. Think about the biographical information about author. The social connections depicted in papers and history books. The cultural leaders of a time, the records of travel, the economic and social influences.

Does anyone know what field of academic study this would fall under?

Time as an asset in mobile development

by

Read on Medium »

On http://ift.tt/1sVNvsn

Time as an asset in mobile development

User experience guidelines for mobile devices have become clearly defined in recent years. Developers have learned how to utilize the limits of a small screen size and prioritize content. Applications targeted at younger demographics (think SnapChat) are successfully leveraging intuitive interfaces for digital natives. There are not many impressive interfaces designed to utilizes time as a resource.
What is out there?

Read later apps touch on the idea of using your current attention to make you more intentional in the future. The notion of a bookmark lets you keep something for later, so you can return to it as needed. Services like Pocket, Instapaper, and Readability, effectively give people a infinite memory bank, where they can collect and store content they discover. This concept works well when used on mobile phones to provide access to content without an internet connection. Similarly, it can make it easy to discover content on one device, then access it on another later.

Time as a design element

Time is an important resource in designing a user experience. Resource value in screen real estate, color priorities, internet connection speed, and many other factors are already kept in mind. Time and its use for the future is not something that actively feel triggered to think about. Time, as a design element, could be used to direct users intelligently.

Current applications are constantly trying to one-up users for a little bit more time. They want to show you one more advertisement or find a hook to get you to come back one more time. Your inbox is bombarded with daily update emails, requests for browser notification access, and too many tracking pixels.

People don’t like distractions. Intentionally creating distractions is disrespecting the end user who makes a company even possible! What if platforms made a better effort at allowing their users to decide how frequently they wanted to engage? What if all the data crunching and analysis that is done to find common trends amongst high value users was used to create a high value experience for users that don’t want to spend their time budgeted for other services?

This is where using time as a design element starts to make sense.
Where it works

Email newsletters are a great to subscribe to a future time commitment for content. Users sign up to email newsletters because they like the content being produced. They are subscribing to the idea that they will get more of the same kind of content. The unspoken contract is that the email address, which is a door into the user’s attention, will be respected.

Email newsletters lose their value when the user’s inbox is overwhelmed with too many emails. Like real life mail, too much email goes unopened and ignored.

Monthly gatherings are another good example. A predetermined amount of time is being committed to a predetermined purpose, with an expectation of some predetermined experience. This is good. For a person who participates in a community, either on or off the internet, a budgeted length of time is great.

The goal here is not to come up with a system for creating a structure for predetermining how time is used. The goal is to come up with a solution for companies resort of attention grabbing.

Where it doesn’t work
The “save for later” services have some basic faults.

First, when the content goes into the bookmarking service, and into the “read later” display, the content producers lose access to the ad impressions/tracking needed to pay for the content. This has its obvious issues for the content producers and seemingly less obvious issues for the consumer. When content producers are not getting a return on investment for their product, they can’t continue producing content. The ads and tracking itself isn’t bad, its the way they are being implemented. Right now, it doesn’t work for the user and the producer.

Second, when the rate of saving content is higher than the rate of consuming content, the saved content is forgotten. People develop habits around bookmarking and categorizing bookmarked content. Most people aren’t sophisticated enough to create a process for consuming saved content, and eventually never do. Saving content without determining when it will be consumed is DOA.

Time as a design resource on mobile

With mobile devices in particular, time is a great design resource. Since user’s are forced to interact with one piece of content at a time, understanding how to leverage the mechanics of intentional delayed engagement is important. Content should be designed with the notion that it is not going to receive the full attention of a user. That doesn’t mean that it can’t at some future time.

What would incentivize companies to adopt time as a design resource?