Marc’s dating dashboard: Tracking Impact in International Development

Since I work in “International Development” I like to use the concept of an “impact dashboard” as a metaphor for getting a handle on my dating life.

Whether you’re talking about solving global poverty or finding the right girl, success or “Impact” is like the weather: everybody talks about it, and nobody is any good at predicting it.

Dashboards are supposed to provide real-time continuous information showing us which direction we are headed. I’ve illustrated the idea below with my own 2012 dating history from OKCupid.

Common Dashboard Features (and limitations)

• continuous flow of quantatitive data and “indicators”
• sparklines
• inputs-outputs-outcomes framework
• visual symbols increase readability and scanning
• the engagement funnel (and on-ramp, retention, drop-off numbers)
• statistical correlations
• benchmarking (see the reference dashboard below it)
• cross-referenced qualitative labels provide context and insight to outcomes
• a bias towards using numbers that are easy to get over those that would best reveal the deeper truths
• incomplete picture – important data points will likely lie upstream and downstream to the numbers your organization is tracking.
• fuzzy math* – if your numbers are exact, they probably aren’t the best  description of the real world in which your organization operates.

*Note: deceptive numbers are not a requirement of dashboards, but they are a feature common to all real world dashboards. In this case, not all 214 conversations are shown, and many of the truly uninteresting data points (“Heheh… Hey baby!” emails) are omitted. For fairness, they are also omitted from the reference dashboard in similar proportions. So we should expect fuzzy math on any real dashboard.

And now… Marc’s 2012 Dating Dashboard:

* Thanks to Dennis Whittle for the suggestion to build this dashboard 🙂

Note that the pink conversation bubbles indicate that the other person wrote that comment in an introductory email (usually the first line). Assume that narrative fragments without pink callouts are my own intro lines. The same is true for D’s dashboard below, but in reverse. All the pink callout bubbles for D are from guys, whereas mine are from women.

I previously posted a detailed analysis of the best (and worst) OKCupid introductions, if you’re curious.

The true utility of my Dating Dashboard came when I benchmarked my activities against someone else who was in the same boat. So below I present a reference dashboard.

And here is my best-matched reference dashboard, from “D”, a girl whom I met through OKCupid and became friends with:

D was very gracious in giving me access to her profile so I could count up emails, responses, replies, etc. I’ve tried to eliminate any personal information here. And since I am not D, I cannot fill in the “feel-o-meter” like I did on my own dashboard. Our outcomes (finding a boyfriend / girlfriend in 2012) were very different.

And a head-to-head comparison of Marc vs D on OKCupid:

As bad as my correlations look (If I initiate a conversation, there is a negative correlation with us eventually going out of a date), my response rate (44% replies) is much higher than the OKCupid average for white guys (29%).

Note that the very start of this process is omitted because I cannot count how many people viewed my profile or how many profiles I viewed in a year. It was probably 600-800 for me. I wrote to 6 women a week – typically 3 women on two nights a week. I realized that women use OKCupid on Sunday and Monday nights especially, so I focused on writing on those days and left it alone on Thursday / Friday nights when nobody uses it.

Observations

•  I had no reservations presenting most of this data, except the sparklines. Somehow drawing a line to represent my emotional rollercoaster experience for each girl felt much more revealing than the rest of the numbers.
• For me – no correlation between any of the stuff I do on the dating site and actual dates. And no positive outcome yet (AKA a girlfriend), though plenty of enjoyable companionship and friendship along the way.
• My best-matched female counterpart has a totally different experience. She is much more cautious and thoughtful about writing to people. For most of 2012 I’d write to 3 new women twice a week, and was 50/50 on ending conversations. She would give many more men a first date and size them up in person. And the men who wrote to her seemed like they needed more screening than the women who wrote to me. But without two dashboards, none of these insights would be possible. Reference data on dashboards is essential.
• Making D’s dashboard revealed some deeper insights to me. Though we’re very similar people, she made the effort to write back to nearly every guy that she dated and give him a wrap-up summary with encouragement and clear reasons for not going forward. “I like closure,” she said, when I asked her about it. I think it is also her nature as a former Chaplain and Seminarian to want to tend to the emotional needs of every person, no matter who they are. I on the other hand, tend to let the mutual silence following a date speak for itself. In the future I will make an effort to wrap up communications with someone that I don’t expect to date again, though it is not a joyful activity – I see that it makes me a better person. (In fact, just this week in starting to write those post-date wrap up messages, I opened myself up to receiving some very terse and blunt “We’re not a match, but good luck in life” messages back. Ugh. But I’ll continue this experiment.)
• Making D’s dashboard helped me walk a mile in her shoes, and see how much harder it is to be the girl on the site. After screening 100+ messages, I am (a) sooooo sure I’m not gay and (b) quite self-satisfied that my messages are way more interesting than the ones she receives from the average guy. I found maybe 1 in 100 guys that wrote her witty, attractive, and intriguing. For the woman on OKCupid, emailing is the beginning of a very long screening process. For the guy, if a woman writes you first, you’re halfway to the date, though my emailing to date conversion rate appears to be 10%, much to my surprise.
• On the selection of numbers – this OKCupid-derived dashboard does what most dashboards do: It selects the numbers that are easiest to get. Although I wished I could correlate match % to dates, I would have to drilled down deeper and none of these numbers are in a database where  I can do that. And that feel-o-meter is what I wished I could track quantitatively, but alas, emotions are not easily quantifiable, so as a proxy I have emails and dates. Also a sparkline of me-vs-her tit-for-tat email patterns would reveal more, but it would take hours to compile for each of the 200 conversations. These are examples of why Impact Dashboards in International Development are so frequently sparse; they go for the easy data, not the revealing data.

How the story ends

About a week after posting this I relocated to Portland, Oregon and went on a date the next day with a girl I’d met over Okcupid. We had little communication prior, and I had no reason it would be any different from the 50 dates I’d had in the past year. It was Christmas Eve 2012, and I am still seeing that girl every day. So no matter how pointless your dates might be, every date you go on refines in your mind one more bit of the puzzle about what you are really looking for.

Postscript

I enjoyed writing this post, and “D” who provided some of the data enjoyed reading it. My mother also reads my blog. She wrote:

“Dear Marc,

Don’t analyze your dating, at least not online, or your % dates are going to be much less than 10%. Love can not be programmed.”

In my defense, we are all searching for meaning. The bigger the questions, the deeper our search should dive. And many of us are curious when, where, and how we will come across the person who will become our life partner. And while I have enjoyed going on all of my dates, I think it would be foolish not to put my full faculties to work in understanding what dating and love is all about.

I was once married, and in a relationship for 9 years, and I’ve written all about it. I think I know a thing or two about love, but that doesn’t mean that I have any advantage in finding it in the next person I meet. The numbers at clear – there is no correlation between activities (emailing, dating) and outcomes (girlfriend), so you might as well enjoy the ride and keep an open mind to every girl you meet. Or study your next date’s ring finger.

Related

Ted: The real algorithm for online dating is something we personalize to what we care about most. No computer will be as good at it as you are.

Infographic: Breakdown of why there are so few quality prospects among black men

How I applied this to international development

First – I decided the outcome:  We (GlobalGiving) seek to provide the means for more money and resources to make it to the higher performing organizations worldwide.

Higher performing can be defined in many ways, but we chose to find organizations that learn from their mistakes and are in the process of growing and improving their work. They are embedded in a community and share ideas with other organizations. They cooperate and collaborate locally and internationally. They listen to their community.

We built a learning and performance dashboard for each organization and an Impact Dashboard that rolls-up all these dashboards into a picture of how well GlobalGiving is doing on our goal.

We now track activities that represent important behaviors:

• Increasing organizational capacity (capacity here means reaching more people, getting more grants, and doing more work more efficiently)
• Learning from mistakes
• Active in a civil society made up for organizations and community members
• Collaborator, cooperative admin style
• Gathering and listening to community feedback
• Growing influence: Increasing access to power brokers; more organizations and community members aware of them

Many of these are “rates” and not traits, but all of them can be measured if we cast a wide net. That is the beginning of building an Impact Dashboard. I am firmly NOT interested in the “McDonalds Impact” approach of counting beneficiaries and adding up the people affected by Development Activities, such as loans or workshops. I call it the McDonalds Impact, because what you end up counting are people whose lives have been affected about as much as if you gave each of them a Big Mac. That’s not changing lives – and yet so many big organizations aim for this number. Acumen bragged in December 2012 of having Impacted the lives of 100 million. They’ve moved a lot of money, but I doubt they’ve got a weekly/monthly dashboard that tracks all of the things I’ve laid out here, and if they did, they’d be posting the kinds of surprising insights it would reveal like I just did for my own dating experiences.