How I become a Data Product Manager Out of College, and didn’t know it.
- Capturing Data
- Capturing Real-time Data
- Reflections on The Ethics: consent & data privacy
Event Farm will always be where I found belonging and a purpose in tech.Robert Azeem Jackson III
Prior coming to the teachers college I worked as a project manager at an Event Technology Start-up in the District of Columbia. Much like our competitors at the time our organization was looking to track RSVPs and “No shows”. Unlike our peers in the event management SaaS space, we were a “big data” organization at our core. We differ from other agencies in a multitude of way, particularly around the methods in which we capture data, the depth of that data mining, and how we get the consent of those in which we capture data.
Our organization primary worked in the business-to-business space, with some commercial application revenue. We would engage in long term contract in which we would send out invitations for private corporate events to our clients stakeholders and employees. For example whenever a large commercial music festival happens in a major city large companies sponsor these events and hold private corporate “parties” and events that are only open to their stakeholders and key movers and shakers.
Traditionally before partnering with our company these events would still happen but the host companies had no way to accurately track that came in and out? How long people were there? When did the largest volume of people come to the private event? Etc. What we promised was to capture that data and other valuable data points that I’ll elaborate on later in the text. The return in investment for our clients was as much about capturing the data points above as much as it was about sample people demographic.
PHASE 1: Method of Data Capture
This all begins with how we capture data [phase 1]. As stated previously we did similar work as companies like Eventbrite, meaning we push e-vites to our client list, which ask them to register for the event using personal identifying information. At the most basic level, their first and last night (which would appear on any guest list) and company they work for. I’ve been apart of projects where at this first phase we ask them more in-depth questions (usually optional, depending on our client’s demands) like age, political affiliations, their favorite cocktail, etc. Between the time in which invites are sent out and the event itself data is capture on attendees. On the an acute level we can learn what time they opened the email, if they RSVP’d how many times they open the email. More obtusely we can learn things like the total number of people who have opened or RSVP’d vs. how many total invitations were sent out. This takes us into the next phase of data collection, which occurs during the event in which people show up and “check-in” to the event or they don’t. At the most simplistic level this is the level of data collection that can occur, but as I mentioned before other organizations already capture this information and this type of data is simply a digitation of traditional paper invites and guest list. Where my previous employer set it apart was in the next phase.
PHASE 2: Depth of Capture
The next phase [phase 2] was the experiential data collection. This refers to the data that is captured of guest while at the event while they are there that goes beyond the survey data we actively consented to giving us while filling out the RSVP. My organization by request of our client, using our experiential technology has been able to capture over 100x the number of data points gathered via survey through the use of proprietary technology and gamificaiton of the guest experience. In this third phase proceeding after the guest is “checked-in” they are then given a credential (a wristband, card, or sticker) that uniquely identifies them.
The RFID/ NFC (new frequency chip) credentials, similar tech as Apple Pay, are needed to interact with the various gamified stations we’ve set up at the event. Whether a guest wants to check-in to VIP, use the “digital photo booth” to upload pictures to Facebook, or order a custom drink that seems to know your pallet they can only do so by using their credential to interact with the various technology stations that facilitate those activities. From first hand experience managing these events I can say that a major part of why guest are there is to have fun and explore the “games” which is why this system works and doesn’t lead to frustration or questioning among guest. To be clear at no point data being captured without the guest engaging with the technology stations. Meaning for the company to capture information on an individual about how many drinks they’ve had they would have had to come up to the drink station device, used their credential to activate the device and order a drink, we would then register this interaction in our database to that credential. Conversely, if that same event occurs but the drink is given to someone else besides the person who ordered it we have no way of knowing that. These interactions between credential and our devices capture 1000s of data points but are also limited to only capturing those specific sequences of actions.
PHASE 3: Rob the Human Algorithm
The final phase [phase 3] is how me as the project manager is responsible for pulling that data from the event from the cloud servers that host the activity from the various station. This data is aggregated and converted to a csv file. The data within this file holds the variable names of the different technology stations as well as the credential IDs but no names. This information must be cross-references with a different set of data that holds the information that was given via survey when the guest RSVP’d to the event. Without this RSVP data I have anatomized data, in the since that every interaction, whether it be a drink order, a game played, or a coat-check, is not assigned to a person but to a credential ID. Only after combining the two data sets can I use an Excel to capture complete stories of individuals using their demographical information.
Longitudinal Reflections on consent then and now (2014-2022)
Reflecting on this 9-month experience
Initial reflection 2016: Having now been 2 years removed from that role there are few takeaways I have begun to shape based on some of the readings; as it relates to my understanding of privacy and consent. In what I describe as the first phase of the work we did nothing about that seemed out of bounds for the type of information needed to run an event, all we did was digitize that experience. Ethically there are no red flags here. Where the ethics of consent become tricker are in the second phase in which guest are expected to blindly accept that the only way to enjoy this event is to work an ID credential and use it at every phase of the event. Full discloser, we don’t spell that out to the guest and that is when things start to get muddy. The 34 CFR in FERPA illustrates in the context of children data where the lines of ethical vs. legal are drawn. Adult data has far less regulation in the US. The reality is the data captured in our events is the property of the client and our company unless otherwise expressed in our contract. These data captured during these events can be very personal. For example if the CEO of a company shows up to a private corporate event and orders 12 drinks using his/ her credential do I have the right to use that info? With respect to edtech, the ability of MOOCs like Knewtown or Khan Academy that can track every aspect of the students learning and offer new ways to study learning is a highly adaptive tool. With applications beyond its initial intent. Data Manager Reflections 2022: When I reflect on these experiences I begin to question the ethical nature of the type of information. Prior I assumed like most technology start-ups that the collection of this data has been vetted by legal teams, but should I as a recent college grad should have had access to the personal information of a senator who comes to a stakeholder event? In theory this info is anatomized by the separation of RSVP data from credential data, both having value separately, but while working at this company I rarely was asked to keep those things separate, I had next to no oversight, and the legal documents we signed with clients had very little information as to what could or couldn’t be done with the info. The data landscape has changed a ton since 2014 when worked as a data product manager; data is a strategic asset, meaning it’s protection, acquisition, and reliability is/ can be paramount to the success of a clients work, and a business. Now, we have to consider how much effort a private company in this case must use to protect confidentiality? Often I was asked in my old role to create data reports for clients as well as infographic use that would go live on our blog. This goes back to the idea of the importance of thinking through the data before releasing it to the public sense we can not monitor use to ensure that the other researcher respects the confidentiality of the respondent, (i.e. how easy would it be for a person to figure out that the CEO ordered 12 drinks based on the information I make public?). Now as a tech founder, former youth Non-profit founder, I think about how private companies that contract with school have evolved the way in which “schools” collect data and the type data they collect.