Strategic Advisor Nora Marketos' interview with Dr. Alexandra Pittman, Founder & CEO of ImpactMapper was originally published in the AI for Philanthropy Newsletter.

Nora Marketos: Today I am sharing an interview with Alexandra Pittman, PhD , the founder and CEO of ImpactMapper . I was intrigued about her learnings in systematically incorporating qualitative data analysis into impact measurement and storytelling. Based on this, she has not only developed a useful product for the grantmaking and social impact sector. Thanks to generative AI, she and her team are now accelerating significantly its potential for large-scale use among a variety of text-based reports and surveys.

‍Thank you, Alexandra, for taking the time to speak with me about your experience. First, I would love to hear a bit more about your background. Why did you start your company, and what excites you about your current work?

Alexandra Pittman: I've been working in the social impact and measurement sector for 20 years. I started my own consulting company in 2005 when I was working on my dissertation. At that time, I started advising a lot of different organizations, such as INGOs like Oxfam, feminist nonprofit organizations, foundations, and UN agencies. I helped them to conduct research, evaluations, and analyses around the impact that they were creating. This is when I started specializing in monitoring, evaluation, and learning (MEL). I have always been interested in tracking change in areas where the impacts are not easily measured by numbers alone, such as women's rights, social justice, or human rights initiatives. Capturing change in this type of work is complex. I love that challenge of crafting evidence-based narratives of change and helping people to truly understand the significance of rights-based and normative change work.

I realized early on in my career that to really understand these dynamic change processes, we needed a much more nuanced strategy for impact measurement than what was taught and being promoted in the international development and philanthropic sectors. We needed to understand the contexts in which the work was located and the pushbacks and leaps forwards that occur in these change trajectories -- that often are decades or generations long in the pursuit of justice and equality.

These pathways are rarely linear, these are usually struggles that are based on deep existing inequalities and systemic discriminations within a country or community. And if funders or organizations want to deploy resources and talent to address these significant inequalities and discriminations, then we need different models for assessing progress than the typical, linear logframe models. So when I helped organizations develop their MEL frameworks, I ended up creating a lot of participatory processes to understand change trajectories and theories of change, and started focusing on qualitative indicators and storytelling, in addition to quantitative indicators of progress.

One of the most pressing issues I saw in my previous consulting business, which still holds true today, was the massive amounts of text data that were sitting around in these organizations’ databases, without ever being analyzed. This was truly amazing information and data on a social change process, strategies that were working, strategies that were not working, outcomes that were occurring, stories of change that completely transformed communities and flipped the analysis and perspectives of progress on its head in a given context. What a huge waste! Data was available, wasn't being used, and yet it was so central.

It felt like my mission was to get organizations to use this valuable text data. So I would help organizations develop outcome taxonomies and then code their rich, qualitative data, and align it with those outcome frameworks. This helped surface incredible insights about their funding portfolios and program work. This allowed me to both quantify the qualitative data as well as keep the nuance and examples of change. I would also collect original data through surveys. When I was running my consultancy firm, I ended up using four different software tools to do my job, NVIVO for the text analysis, SPSS or Stata, collecting data in Survey Monkey, then I'd visualize the data in Tableau. For 10 years, I used all these different software tools to do my job.

‍While I love data analysis and data crunching, it was super apparent that the average program officer in a foundation wasn't going to have the time and capacity to use all of these software tools for their impact reporting. They would just rely on the existing software products out there, which all focused on quantitative data, which meant that they would throw the qualitative data out the window.

So much funding goes to what is measured, and if the existing software tools only measure quantitative data systematically, then we are losing out on understanding some of the most impactful solutions based on contextual, qualitative data, stories, and insights, because you've immediately excluded their relevance, right?

‍I realized that this was a huge issue in the sector that I wanted to solve.

I concluded that to really have a robust system and structure for impact data analysis, we needed a software tool that integrates qualitative data, quantitative data and financial data in one place, combined with data visualizations. And that was the birth of ImpactMapper in 2015! I used small bits of my consulting money and bootstrapped working with developers to develop this tool and have received foundation grants and seed equity investment to grow. We've been growing and building ever since, with foundations and nonprofit organizations, evaluators, and researchers as clients and partners in defining our features and product roadmap. ImpactMapper is a robust mixed method impact analysis and reporting tool, and I can talk more about its features in a minute. It’s been an exciting pathway and journey. We’re constantly upping the game in terms of being able to serve the philanthropic and nonprofit sectors and move money to what matters. And to move it in strategic ways, through analysis, data, and learning.

NM: Can you share more about what the ImpactMapper software does?

AP: Our product is suitable for many different types of organizations. A lot of funders use our tool, including foundations, impact investors, multilaterals or UN agencies, and also nonprofits, researchers, and corporations. We've built the software tool to be very flexible in its analysis capabilities and use cases.

If you are a funder, you can upload all your grant information on ImpactMapper and visualize your grantmaking trends over time. We have APIs with Fluxx, so you can pull that data automatically over or if you have grant data in an Excel file, or in another grants management product you can export it in Excel/csv and upload it into ImpactMapper. This year we are building APIs with Airtable, Submittable, Salesforce, and an integration with Giving Data.

Any user can upload qualitative data, reports, or stories in any PDF or Word document they have into the system, align it with your custom taxonomy and aggregate and visualize those trends through our robust charting tool. The taxonomy could be your Theory of Change, outcomes, or a strategic plan. You can also create a taxonomy from the data itself and do inductive coding analysis on any text data you have. Additionally, there's often numerical or financial data embedded in text documents that you can pull out, aggregate, and sum up to be charted. You can also save quotes and stories of impact to bring your analysis to life.

We also have a custom survey tool to collect original data. It has a lot of amazing features-- a couple of unique features that differentiate us from other products include the ability to code open-ended survey responses. You can tag the data like you would in the text analysis tool or you can have the participant who filled the survey out, tag their response for you. This means ImpactMapper has a participatory coding tool embedded, which I think is amazing. You have instant visualization of all the survey responses they roll in, and you can do some very cool qualitative data analysis and save quotes on the surveys as well. The whole tool is built to be mixed methods for the purposes of research, evaluation, storytelling, or impact tracking. Of course, ImpactMapper has a really strong focus on lifting up and analyzing more systematically qualitative data, that is what differentiates us, because that is what's missing in the impact measurement software sector.

NM: Thanks so much for sharing your amazing journey. Based on my experience as a Foundation Program Officer I can just confirm that usually you don’t have the software tools that grasp these qualitative nuances nor the capacity to do such analysis on your own. I know that you have made a huge advancement by also integrating generative AI in your product. Tell me a bit more about what your basic offer is, and when and why you decided to integrate data analytics and Gen AI.

AP: In 2022, we engaged with equitable AI in a pilot with Malala Fund.  We started with a sample of their grantee reports, used their existing outcomes portfolio framework for the Education Champion Network, and started to train datasets and developed an autocoding model that could be applied to the text data. It would pick out sentences that looked relevant to a given outcome that Malala Fund cared about and suggest a tag. In a second step, the users could review this and see a data visualization. We were lucky in this case because the previous year we had also manually coded a set of these grantee reports so we had a training database to start from. We saw some really interesting results. However, there was great room for model improvement and reliability gains.

Then Generative AI came out on the public scene with the release of ChatGPT. We've been very thoughtful about this process, because we are engaged in and support the responsible use of AI. We've been thinking through this a lot and I love to speak on this subject in philanthropy. We decided there were too many potential security issues with ChatGPT. So we have been building a solution that keeps information encrypted and within our database, while leveraging LLMs, such as Claude. No personal data is shared back to training for any other product or services, which we're really proud of.

As of now, the first two public AI features that we released in December 2024 are connected to the survey tool itself. The first is a huge time and money saver for organizations that are working internationally and across different languages. We just launched an automatic tool that allows you to translate a survey into pretty much any language with a click of a button. This is really exciting, as it takes that first cut of a translator out of the equation and saves valuable time and money. You still need to engage a translator in the review. We’ve built in tools to make that review process very easy for the user and the translator, so they know what has been reviewed by someone and accepted, and what has not. This feature allows for survey data to be submitted in one’s own first language, so this is culturally sensitive and supportive. We're thrilled with this feature!

The second feature that we also launched in December is called AI Insights. The feature gives you a summary and a short analysis of any survey question and chart that you choose by pressing a button. This helps you to move faster along in the data analysis and writing process. Again, as with any AI, we need to have a human review this. We always support a human in the loop and we make how AI is used in our tool transparent. This is something that we share with all of our clients: you have to have a human in there to fact-check. We know these models are hallucinating and have biases. It’s not a copy-paste, move on scenario.

We are now getting back to that original project and are working on the big, hard problem of structuring and autocoding unstructured text data. We have an incredible foundation with the manual tagging we already have built into ImpactMaper and our learnings from the Malala Fund pilot. This work we are now engaging in will be a game changer for philanthropy and a significant dream I have had since I started ImpactMapper. This is an exciting project and we are looking for partners who also are interested in this journey by providing their grantee report data and insights and support.

Another thing that we are really excited about, that I mentioned earlier, is giving users access over if and when they use AI in our software. We’ve built the AI features and functionalities to be very explicit and transparent. We are not using background processes like some other software companies are, to analyze every data point that you have in your system. You need to opt in to using GenAI. There is a price on that of course. We have you push buttons to say, enable auto translation on surveys. If you don't press that button, there's no AI on your data and your account. With the auto-tagging tool that we're developing, it's the same. There's a button that you have to press. So we want users to really start being aware of the GenAI that they are using. This is because there's so many shady practices out there, where you're not sure how and when AI is being used in a software product you’re on and if you were automatically opted into sharing your data for training. Especially in this space of philanthropy and impact, we are talking about sensitive data, and we need to be thoughtful about it as users and consumers of software and protect our data.

NM: To conclude, I wanted to get your reflections on what the key requirements for foundations to engage in such a process are. What needs to be in place so that foundations from all sizes can fully harness the potential of genAI?

AP: Foundations are already sitting on a significant amount of information and data that can be leveraged. So the next step is to identify parts of the work that could be valuable to automate or gain deeper information on, identify different pilots that could be taken, and hopefully build products that grantees can co-develop or inform, all while being mindful of data security and privacy.

There are so many exciting horizons where AI can help us to optimize the understanding of what's happening in the world, how to move money more effectively and more efficiently, and to really make a bigger impact and learn from the data that we already have. I think one of the first places of focus that is a no-brainer is grantee applications. This means using qualitative data analysis tools to start to code grant descriptions and other data and align it with custom taxonomies that a foundation has. There's a lot of information that can be mined around socio-political trends. In my view, we're not really optimizing this data and the context analysis lying latent in grantee applications right now.

As a sector, we should think more about the potential of what can be achieved when we start to collate these trends, share them across funders and nonprofits, and fully fund the most underfunded and significant issues of our time.

Cross-sector learning and collaboration around grant and contextual demands per thematic, issue, or sectors, and thinking at a systems level could also be transformational. So, I think the sector definitely can start moving there with coordination at the impact level. There’s just so much data in software tools at donor and nonprofit institutions, but it is not even being used.  Of course impact reporting also offers a huge area of potential— to really understand what strategies are working at scale, what's not, what are the trends in outcomes across regions and sectors, etc., where did we make mistakes that we need to learn from, and that others can learn from. And how do we share this data more strategically, so others don't fall into the same trap. So, AI can do this qualitative analysis well, and investing in it can yield incredible results when you look at this at scale.

Another area that we've been working on quite heavily over the years is evidence synthesis. We do this with a lot of UN agencies and Spotlight Initiative, where there’s tons of evaluations that have been published over the years, research that's been published about lessons learned and case studies of success that can be actually aggregated up to synthesize and share the insights.

Honestly, this is a really exciting time to start leveraging these insights and move money more strategically in the future. There are a lot of amazing insights to discover through better use of our data.

NM: Thank you so much for that more visionary, high level reflection to conclude our exchange. It's super exciting to hear about your journey and targeted offer for the funding and non-profit community.

More About ImpactMapper​

​ImpactMapper offers a 360° set of services, ranging from impact metrics and management, evaluation reporting, training and capacity building, and proprietary software to ensure groups are on track at every stage in the impact tracking process.

We work with organizations that value human rights, equality, and social and environmental justice and place these values at the center of their work. ImpactMapper is a trusted partner, working with some of the largest brands, including UNDP, Spotlight Initiative, Winthrop Rockefeller Foundation, Malala Fund, Porticus, IKEA, DLA Piper Law Firm, Women Deliver, Norwegian Human Rights Fund, ICAN, Chanel Foundation, Cartier Women’s Initiative, and MIT Data and Feminism Lab to name a few.

Our proprietary technology and software platform is unique in the philanthropic and investment ecosystem as it allows for deep analysis of qualitative, quantitative, and financial data in one tool along with visualization. Our competitive advantage is in transforming text data into quantitative aggregate insights to be analyzed along with financial and other custom quantitative metrics, KPIs, and the SDGs, which helps our clients tell their full impact story.

If you are interested in learning more, using our tool or supporting ImpactMapper on the path to transform social impact tracking and collaborate in our development of AI-driven impact reporting infrastructure, then please reach out. alex@impactmapper.com

‍