M+R Benchmarks Webinar

This presentation compared nonprofit marketing data year-over-year from 2014,2015,2016. They prefaced the talk with two high-impact points:

  • M+R’s data team is also their creative team. They blend data analysis with visual thinking early and often
  • Benchmarks are just that. They are clear not to provide any “WHY?” suppositions along with their findings. They are comparing hundreds of nonprofits and measuring things like “number of emails sent per year”, and the purpose is more or less “Am I normal?” and not much more than that.

Although it’s interesting, that second point. The presenter kept making suppositions! I guess it’s hard not to try to tell a story with data.

Some of the data they found that stood out to me:

  • .38% click-through rates on fundraising emails
  • .05% response rates on fundraising emails
    • cultural orgs do a little better
  • $62 earned revenue/1,000 emails delivered for cultural organizations
  • Online $ increased by 14% in 2016
  • lots of nonprofits are performing well with monthly giving, but cultural orgs just aren’t doing it
  • .3% website donation conversion rate
  • 22% conversion rate from the DONATE HERE page
  • How do we crack the code on what is important to measure in social media?
  • Facebook – the people who are engaging with your FB are usually not your fans.
    • Most of the people who see your posts are not your people. Think about this when designing content to be new-user-friendly. (Eg. “If you like this kind of post, click like”)





Stat Trek Captain’s Log: Lesson 1

Finally getting started (formally) learning statistics, by going through the AP Study Guide on Stat Trek. So good! Today I learned something I don’t think I ever realized about mean vs. medians:

  • The median may be a better indicator of the most typical value if a set of scores has an outlier. An outlier is an extreme value that differs greatly from other values.
  • However, when the sample size is large and does not include outliers, the mean score usually provides a better measure of central tendency.


range – the difference between the largest and smallest values in a set of values. In 2,4,9,5,7,3   the range is 9-2=.

pros: easy to calculate
cons: ignores middle values


interquartile range – divided an ordered data set into four equal parts.

first, arrange the data set in numerical order

Q2 – median of the entire data set

Q1 – median of the data below Q2

Q3 – median of the data above Q2

IQR = Q3-Q1


Q2 = (4+5)/2 = 4.5




pros: not strongly influenced by outliers


variance – average squared deviation from the mean.
variance of population= σ2 = Σ ( Xi – μ )2 / N
variance of sample= σ2 = Σ ( Xi – μ )2 / (N-1)

σ2 = variance
Σ =summation
Xi = element i from the data set
μ = mean of the data set
N = number of items in the data set

Σ ( Xi – μ )2 = deviation sum of squares

example: 0,1,5,6
deviation sum of squares= (0-3)^2 +(1-3)^2 + (5-3)^2 + (6-3)^2

variation = 26/4 =6.5


standard deviation – is the square root of the variance! super easy if you understand variance. The standard deviation is a numerical value used to indicate how widely individuals in a group vary. If individual observations vary greatly from the group mean, the standard deviation is big; and vice versa.

A standard score (aka, a z-score) indicates how many standard deviations an element is from the mean. A standard score can be calculated from the following formula.

z = (X – μ) / σ

where z is the z-score, X is the value of the element, μ is the mean of the population, and σ is the standard deviation.

Here is how to interpret z-scores.

  • A z-score less than 0 represents an element less than the mean.
  • A z-score greater than 0 represents an element greater than the mean.
  • A z-score equal to 0 represents an element equal to the mean.
  • A z-score equal to 1 represents an element that is 1 standard deviation greater than the mean; a z-score equal to 2, 2 standard deviations greater than the mean; etc.
  • A z-score equal to -1 represents an element that is 1 standard deviation less than the mean; a z-score equal to -2, 2 standard deviations less than the mean; etc.

Breakfast Club

This morning I went to a Breakfast Club chat at General Assembly, where they served coffee and bagels, and hosted a little chat by Lilibeth Gangas, Chief Technology Community Officer at Kapor Center. From her staff bio:
In this role, Lili helps catalyze Oakland’s emergence as a social impact hub of tech done right – where tech, diverse talent, and action driven partnerships can tackle pressing social and economic inequities of our communities head-on.
She talked about her winding career path and a few of the questions that have guided her work, like “How can I have a positive impact on the people of the world?” and “How can I best use the talents I’ve been given?” The biggest lightbulb moment for me was when she mentioned that moving from the for-profit world to the non-profit world, was challenging at first, because she had been so used to working to meet clear ROI’s, and the non-profit world is less business-oriented. During the Q&A I asked her to elaborate on this with the question, “So, in the nonprofit world, how do you measure success?”
In her response she talked about the pitfalls of over-quantifying, making up metrics for metrics sake. Best business practices say that you should never be focused on more than 5 metrics at a time.
She talked about how Kapor Center is “lucky to have PhD’s working with Theory of Change” who break down the big lofty goals that drive the nonprofit into smaller strategies, tactics, and measurable deltas. Of course they’re looking at things like “how many people did we place into jobs”, but it’s not just direct impact that you have at a nonprofit. You have to think about the indirect impact as well.
She ended with simple, if mystifying, advice, “Less is more.”
I’m thinking I want to reach out to her and some of her colleagues to talk about their Theory of Change (this is something Nina Simon mentioned during her CAM talk as well), and how we might be able to adapt this to the art/social justice we’re working on here. I feel like we can measure sales and donations, we can track attendance, demographics, and psychographics. But we’re missing something. We’re missing that intangible inspiration that happens outside our walls, maybe takes 10 years to become visible, that spreads and makes culture subtly slowly surely shift.

Don’t Let Imperfection Immobilize You

At MGI we often hear that “IT has the data, and it’s a mess!” Well, here’s a dirty inside secret…almost everyone’s data is a mess. Make a start. Figure out how to get the best set of reports you can, task the organization with monthly procedures to assure you get it, and then don’t worry too much about the rest. As long as you know your KPI’s, you can stay ahead of your game. Improvement will follow.

Marketing General Incorporated — a company that focuses on helping companies grow their membership programs — has a few tips about how to work with your data:

  1. set clearly-define KPIs. They sited the following ones to get started:
    1. Response rate to campaigns
    2. Cost to obtain a member
    3. Renewal rate
    4. Average tenure of membership
    5. Lifetime value – the value of a member over the time of their membership
  2. establish data rules – but don’t become paralyzed by making your system perfect. just get the groundrules down, and then get started. see opening quote.
  3. actually use your KPI’s to drive planning, budgeting, and success


I’m at the end of my first week in my Data Analyst role, and the main project I’ve been working on is helping to define KPI’s for a new-and-improved dashboard for next fiscal year. The complaints with the old dashboard included it was time-consuming to gather the data, and no one was confident that the points being tracked were the right ones. It also seemed to me to be clunky, kind of hard to read, and not clearly actionable.

So we decided to make the metrics hard to dispute — we’re picking things to measure that directly tie into the strategies outlined by our most recent strategic plan, so that this dashboard acts as a progress report on where we are in terms of this strategic plan.

So it’s interesting for me to read about these data methods that MGI outlines. It helps to keep me grounded, since the more I dig into this process — interviewing key staff members about how to measure the success of their programs that are called out in the strategic plan, building out my recommendations — the heavier it feels. This is a good reminder to keep it simple, and fight my perfectionist tendencies to go down data rabbit holes forever.

Audience Segmentation with Howard Levine at the 92nd Street Y

This film is a few years old, but I’m loving this story on how an arts org started using audience segments to create better targeted messaging. At my organization, we have our target audience segments defined, but we haven’t started *using* them like this yet. Great inspiration.

Do you truly understand what makes your donors and patrons return, or not?

Traditional, transactional arts data — recency, frequency, and monetary —  don’t truly provide enough information to drive future marketing and programming decisions. Solely relying on this info, you’re missing half the picture. Why did you come, what drove you to participate or donate?

Lessons we can learn from the financial industry:

  • past performance is not an indicator of future returns
  • what leads to market changes is more often due to emotions, not statistics

Arts intrinsically elicit emotions, but how do those emotions drive people to attend and donate? You have to look at their beliefs and values.

92Y segmented their audience based on values and beliefs, using the Morris Hargreaves Macintyre cultural segmentation system. Then they overlayed these attributes with the ticket and contribution history in their databse.


  1. Create baseline understanding of their audience and their position in the context of the larger NYC cultural market. How did they compare
  2. Deep dive into motivations, values, and needs of their audience segments, to better reach, engage, and develop those audience segments
  3. Institutionalize this learning — make segmentation standard across marketing and fundraising

This is how they did it:


Now every conversation they have, from content planning to fundraising, includes these segments — which of them are they trying to message to and how can they tailor that message.

Stanford PhD Data

I’m lucky to live in SF while trying to figure out this data world, since I, as a result, happen to be surrounded by data people and programmers. My roommate is doing phD work in creating an algorithm that predicts the success and failure of medical treatments by aggregating data outside of clinical studies. Looking just at patient data across multiple hospitals, he can run tests on the likeliness of outcomes. This saves time/money by working outside of the clinical studies model and harnessing data that’s already there. This is the kind of outside the box thinking I am going to be needing to harness from now on. 
How do you measure impact? Inspiration? Inclusivity? How do youbtell the story of change and progress over time? What lies beyond attendance numbers and ticket sales? How do you successfully measure incubation, reach? What story can you tell to make the arts feel just as critical in the world as any other poltical or social cause? Where and what are these numbers?

Blog at WordPress.com.

Up ↑