Case Studies

The problem/situation: When I first joined the Girl Scouts they did not yet view data as a strategic asset, and there was no one with a holistic view of their data practices across both national headquarters and 112 local councils. As such, they lacked the solid foundation needed to fuel their plans for full-scale digital transformation.

Obstacles/challenges: Individual business units were managing (or not managing) their data separately, as well as making siloed decisions about their systems. Nothing was interconnected, which was costing the organization both in terms of unnecessary expenditures and in the number of key business questions they simply could not answer.

My team’s approach: I met with stakeholders across the organization to understand their business challenges and how they were meeting them within their domain. Leveraging my natural ability to see interconnectedness where others see complexity, I developed a multi-year enterprise data strategy that laid out a unified approach to integrating current and future data needs from across individual departments and councils.

Results/impact: In a few short years my team and I created the critical data infrastructure necessary to fuel the Girl Scouts’s digital transformation, and the tools needed to implement data-driven decision making moving forward. As part of creating this infrastructure, we:

• Reversed widespread data loss to dramatically increase the amount of data available to the heads of individual business units and local councils; Established national data health metrics; implemented best in class tools;

• Streamlined our processes around data collection and reporting;

• Moved from relying on numerous consultants to fostering in house talent;

• Delivered a new enterprise data warehouse and new business intelligence and analytics tools;

• Rolled out a comprehensive literacy and training program to build analytical muscle across the national and local Girl Scouts system;, and

• Established a data governance program to achieve network alignment around key data concepts such as data ownership and privacy.

The problem/situation: An estimated 50M+ American women have participated in the Girl Scouts — but the organization’s disorganized data storage and lack of clear data strategy made it difficult and, sometimes, impossible for staff to contact or activate this potentially powerful group of supporters.

Obstacles/challenges: Girl Scouts used to wipe its national databases every few years due to the previously high cost of storing data(something big data has eliminated.) As a result, staff had only identified 300,000 Girl Scout alumnae across a number of sources (compared to the estimated 50M+ American women who have participated in the Girl Scouts), and every year another 1M+ alumnae were at danger of being forgotten.

My team’s approach: After analyzing the situation I realized that the primary membership database was configured to only contain current customers and their transactions. Any former customers (such as former girl members who were now 18+) were removed after 2 years, and there was nowhere to house their information. My team developed an approach to automatically “graduate” all former members who had turned 18 that year into adult alumnae records and made them available via our new data warehouse as alumnae. If they chose to re-engage with the Girl Scouts over time (such as having a child who became a Girl Scout), our new data processes would match them to their former record.

Results/impact: We grew our initial alumnae list of 300k to over 9M and growing. By cleansing and enriching the data, we created an audience for digital outreach (increased engagement, increased donations, etc.), and a new data set for deeper analysis and longitudinal research on the impact of scouting in a girl’s life.

The problem/situation: Girl Scouts was at the beginning of a digital transformation, which included building brand new digital products for many long-standing business programs —including the award-winning Digital Cookie platform, which delivered the first multi-channel e-commerce platform for girls participating in the organization’s beloved cookie program.

Obstacles/challenges: Within days of the initial launch of Digital Cookie, the organization became inundated with customer service calls and emails from frustrated parents and customers. The volume was significant enough that the organization created and staffed an emergency help desk, which cost an unbudgeted sum in the high six figures. We (literally) couldn’t afford to have a repeat of that in year two.

My team’s approach: After analyzing the customer support tickets, we discovered that over 80% of the initial customer service issues were related to poor data quality. Simple issues like missing or incorrect contact information and girl date of birth could result in some girls never being invited to set up their online cookie shop, or finding too stringent controls on what they could do with their shop without parental intervention. My team developed a plan to address the issue head-on and well in advance of the targeted launch date for year two. We started by assessing data quality across all 112 councils for the core set of customer data required for Digital Cookie. We then set national benchmarks for data cleanliness that councils needed to commit to in order to be allowed to launch Digital Cookie in year two. We established a master data quality dashboard that we made available in real time to councils, allowing them to track their progress against national benchmarks. We also launched a data quality education program for councils where, during weekly webinars, we taught councils how to read their reports and more importantly, how to remediate their data quality issues one by one. We provided additional one-on-one support for those councils with the largest data quality gaps to close, and stepped in to run automation on their behalf when necessary.

Results/impact: We drove a 400% improvement in enterprise data quality over six months, with all participating councils hitting their benchmarks in order to participate in year two of Digital cookie. By eliminating the data quality issues, we ensured that the volume of customer support tickets remained manageable. We delivered these results for a fraction of the cost spent the previous year on emergency customer support.

The problem/situation: Staff at all levels across the organization, at both national and local offices had a shared problem. No one had access to accurate, real time data to inform their decision-making. Most analysis was done manually via Excel, and recreated every time the data needed to be updated. Confusion could arise when numbers didn’t match across reports or across time periods, as no data lineage was available. Leaders were stuck acting on old data, without a clear sense of what was accurate. Too many questions simply could not be answered based on the data and tools available.

Obstacles/challenges: The existing national data warehouse and reporting tools were outdated and hard to use. The data warehouse only contained current year data and a snapshot of the previous year’s summary data for comparison. The business intelligence tool was outdated and hard to customize, requiring the national office to build nearly all reporting manually. Only 900 staff members out of thousands across the country used the reporting tool more than once a year. As a result, thousands of hours of effort were being wasted each year as staff labored to manually answer core business questions.

My team’s approach: We took a two-pronged approach to tackling this challenge - updated tools and a data literacy program. We started by first making sure we understood our customers (staff, this case) and their needs. We issued a survey to see what people liked and didn’t like about the existing systems, and to hear what they felt were important criteria for any replacement systems. Using that criteria, we found and implemented a new business intelligence (BI) tool that was modern and much easier to use. Underneath the new BI platform, we re-architected the data warehouse, going back to the original data sources to provide new access to 14 years of rich, managed and curated data for analysis. Rolling out these tools, we introduced a new ongoing data literacy program to train new analysts in the fundamentals of solid data analysis and custom report generation.

Results/impact: The results are in the numbers: Our active user base grew from 900 to 17,000+ over a six-month period. Our new national suite of official dashboards and reports eliminated the manual effort previously needed to answer core business questions and leaders at smaller councils with more limited staff resources now had the same access to high quality analysis as their larger counterparts. With our newly expanded data warehouse, the organization was finally able to start tackling business questions we’d never had the data to answer.

The problem/situation: Girl Scouts lacked a unified approach to making decisions and creating policies around data.

Obstacles/challenges: The organization was heavily siloed, with individual business units and councils heads down in their own priorities and challenges. Most did not understand how their data collection or usage impacted others, let alone the customer. (As an example, customers routinely received a barrage of emails from all different units of the organization - none of whom were aware of how much the organization was emailing overall.)

My team’s approach: I stood up a Data Governance council which included executives from each of the major business areas and several Council representatives to make collective decisions related to data usage, sharing, design and management. The group reviewed, discussed and approved my proposal to adopt a ‘customer centric’ approach as our core tenet. This required a seismic shift from each individual unit asking ‘what does my team want from the customer’ to instead asking ‘what is the customer’s expectation of how we would handle this?’

Results/impact: As a result, Girl Scouts was able to dramatically streamline our systems and reduce the previously overwhelming complexity of our security and privacy settings with our core products, which had been inhibiting their expansion/future development. Staff across the organization were better positioned to support our customers, and customers were given substantially more control over how they chose to hear from and engage with the organization.

The problem/situation: When the Covid pandemic hit the country in 2020, Girl Scouts saw an immediate disruption of their entire business. The core Girl Scout experience is centered around in-person meetings. When schools closed down and in-person contact was actively limited to avoid exposure, the organization found itself needing to immediately pivot to digital experiences.

Obstacles/challenges: As the core in-person meetings stalled, all related revenue immediately dried up as well. (After all, who needs to buy uniforms and badges, or feels prepared to renew their membership, in the middle of lockdown?) In order to develop new business strategies quickly, we needed information to help us assess and forecast the impact to our revenue.

My team’s approach: My team immediately brought in COVID positive cases data by county and compared it against our membership, producing heat maps of impact across the country that could be drilled into for county-level comparisons. We then began a prototype machine learning approach to integrating our sales data and active membership data at an individual customer level — something we’d never been able to do previously — to determine who was still making online purchases, and finding similar audiences to target to drive sales.

Results/impact: By mapping this data and refreshing the sources daily, our national office and local council staff were able to see the high-priority impact areas where membership renewal and new online activities needed to be focused. By using data science techniques to target key consumers, we enabled the organization to maximize revenue from the online store.

The problem/situation: During the financial crisis of 2007 large charities were struggling to raise funds at their annual benefits. The first two large galas of since the start of the recession had disastrous fundraising totals at the end of the night, putting the rest of the philanthropic community into a panic. At Robin Hood, the concern was justified: Nearly 100% of their donors worked in the financial field and it was clear that the tactics that had worked in previous years would not match the current challenge.

Obstacles/challenges: Robin Hood’s annual benefit — the largest charitable gala in the world — was just over two months away. Any changes in our approach required immediate action in order to be ready in time. The leadership team decided to introduce anonymous bidding at the event so that our ultra high net worth guests could comfortably donate without drawing attention to themselves by doing so publicly; this would require equipping each attendee with a device that would allow them to make their bids. As far as we could tell, this had never been done before at the scale of the Robin Hood benefit. It was up to me to figure out how to make it all happen. One major challenge was how to find 4,000 identical devices that we could guarantee would work in the notorious Internet “dead zone” of the Javits Center (the only space large enough in Manhattan to accommodate a 4,000+ seated dinner).

My team’s approach: After exploring and testing our options, we selected a vendor that specialized in technology for shareholder meetings. I put together a game plan for the evening using their software, infrastructure, and devices, and began our rigorous planning and testing process to ensure the evening was a success. On the night of the gala I trained and led 250 volunteers to roll out a new electronic check-in process, and to provide high level support to our VIP guests during the evening.

Results/impact: As a result of our work, the organization raised a record $89M in one night at our annual benefit — at the time, the largest one night donation tally for any charitable event ever. The next day I began fielding calls from other charities across the country and the world, wanting to know exactly how we did it.

The problem/situation: Following the success of the previous year’s gala, it was clear that the smart application of technology could deliver outsized impact to the organization. I saw that here was further opportunity to leverage data and technology to evolve the event’s effectiveness — both for the organization and for our guests.

As amazing as it may seem now, back then, much of the world’s largest and most successful charitable event was heavily paper-based, from the registration and check-in processes to the way guests found their seats. The organization only collected contact information from those who purchased tables at the benefit (at $250k each), leaving the foundation with no visibility into the 4,0000 guests in attendance, or any way to reach out to them after the gala. Because the registration process was paper-based, any changes to the guest list were halted early on the day of the event in order to leave enough time for paper copies of the full guest list to be rush-printed and delivered in time for check-in. As a result, guests waited in long lines (sometimes as long as 45 minutes) in order to gain entrance to the event. It was clear to me that taking these processes digital would eliminate these issues. We needed a data strategy!

Obstacles/challenges: The biggest obstacle to making more of the processes around the gala digital was a strong resistance to change from the team running the registration process. While truly exceptional at their jobs, they felt that given the event’s history of success, making any changes to how the event ran meant introducing unnecessary risk. To earn their buy-in, we needed to help them understand just how much risk sticking with a paper-based system represented, including illustrating downstream impacts.

My team’s approach: Understanding our stakeholders’ strong resistance to change, we took it slow and held a series of meetings to walk through our proposed digital process, stopping to address each question or objection raised. We then problem-solved together on the spot to figure out an acceptable solution, and as their objections began to wane, they began to see for the first time how the new process would benefit them as well. They became cautiously excited.

By establishing new data collection standards to ensure we would have full contact details for all guests at the gala, we were able to start front loading guest information into a new, centralized event registration database well in advance of the benefit, minimizing last-minute scrambling on the day of the event. We were able to forgo paper lists at check-in (though we still printed a copy for the team’s peace of mind), and were able to continue making last minute changes to the guest list even after the event had started, which greatly pleased our table buyers. Because every check-in station now had a full copy of the guest list electronically, we were able to completely eliminate the long lines that had segmented people in alphabetical order, which also eliminated the long lines that had become an unfortunate signature feature of our event. .

Results/impact: The contrast between the typical customer experience at check-in, and the streamlined digital experience that my team created, was like night and day. Rather than waiting in line for up to 45 minutes, guests were now on their way inside the venue for cocktails in under a minute. What’s more, the organization now had a complete electronic guest list available within minutes of the event ending, which we could then segment and load into our email platform — putting an end to the all-nighters staff had pulled in previous years to manually make this happen. By putting donation devices into the hands of all guests we drove a 4000% increase in the total number of donations at the benefit, and could now follow-up and cultivate these donors for future donations and events. The registration team was so delighted by the results that they fully adopted the electronic system in subsequent years, even hiring a new team member with more technical know-how to help continue to digitize their processes.

The problem/situation: While at Yale I led the second largest prospect research and analytics team in the United States, supporting the university’s $3.5B capital campaign. The analysts were exceptionally talented, producing thoughtful, detailed research for gift officers across the university. Our work product was excellent, and yet customers were still sometimes frustrated with the results, feeling like they received a wealth of information but were still left without answers to questions they felt were key. Staff felt overwhelmed by a seemingly never ending queue of requests, and yet managers, who had no insight into the work being done for their teams, felt they were never getting enough of our research time. We needed a data strategy to enable visibility and insight.

Obstacles/challenges: The first obstacle was technological. No centralized tracking system existed, and all incoming requests of analysts were being handled individually via email. We had no visibility into the volume of requests coming in or how work was distributed across customer departments, nor did we have high level insight into the kinds of questions being asked. Every request received the same prioritization and level of effort.

My team’s approach: We first implemented a centralized request tracking system directly within the donor database and made it fully transparent and available to all users. Customers could now see the full volume of requests, and understand the competing requests for our analysts’ time. We coded each type of question, and retroactively added as many previous requests as we could into the tracking system, giving ourselves the ability to track and analyze our work over time.

Results/impact: Using this new data, we were able to analyze the kinds of requests being made and introduce new types of research products. By allowing users to ask smaller, more targeted questions that could be answered in X amount of time, rather than always requesting a full research profile (8-12 hours of work), our queues began to shrink, and customers were happier. We introduced new monthly reporting that we shared with each manager, allowing them to see both the number and types of requests their teams had made in the previous year and year to date. Most were surprised to see just how much support their team was receiving from us, especially as a percentage of our overall availability. Leaders now had the ability to work with their teams to make the most effective use of their research hours. Using the data we were now collecting, we were also able to start demonstrating the direct impact of our work on donations (something that had always been elusive), creating reporting that tied our research effort to new gifts being made.