Heath Johnson Consulting

Tweet 1: Most nonprofits don't lose donors at the ask. They lose them in the gaps. The thank-you that arrived six weeks late. The mid-year touchpoint that never happened. The renewal email sent to someone who already lapsed. The ask is usually fine. The work around the ask is what breaks. Tweet 2: When I mapped out the fundraising operation at The Matthews, I found something uncomfortable. Most of it isn't relationship work. Most of it is administrative processing — logging gifts, triggering acknowledgments, flagging renewals, tracking benefit fulfillment. None of that is the relationship. All of it has to happen for the relationship to survive. Tweet 3: Two types of work. Completely different requirements. Relationship moments: the personal thank-you call, the face-to-face ask, the moment a donor tells you what the organization means to them. These cannot be automated. The moment you try, the donor notices — and it's worse than silence. Administrative processing: everything else. Tweet 4: The failure mode in small nonprofits is that both types of work live in the same person's head, undifferentiated. When that person gets pulled into a crisis — and in a three-person shop, the crisis is always arriving — the processing stops. The gaps open. The donor doesn't notice immediately. But in six months, the relationship has cooled and they can't say why. Tweet 5: Most CRM-driven fundraising gives you a firehose. Here's everything that's due. You sort through it. The cognitive overhead of the sort becomes its own bottleneck. You end up triaging the triage. What you actually need: three to five specific actions, scheduled to specific days, with context attached. Not a firehose. A calendar. Tweet 6: This design principle — separating relationship moments from administrative processing, then systematizing the processing — is valuable before you touch a single piece of software. If you can't describe the processing steps clearly enough to write them down, you don't have a technology problem. You have a clarity problem. And technology will make it worse. Tweet 7: Here's the exercise worth doing before your next major gifts push: List the last three donors who lapsed — not the ones with a reason to stop, the ones who just quietly stopped. What happened in the gaps? Usually nothing. The processing slipped. The relationship cooled. Nobody noticed until the renewal didn't come back. That's an administrative failure. Not a fundraising failure. And administrative failures are fixable.

Tweet 1: Candid charges $3,500 a year for access to public IRS data. Every nonprofit 990 is public record. Free. Queryable through a public API. They wrapped a search box around it. Tweet 2: Small nonprofits run grant development on board connections, Google searches, and whatever the previous ED knew before they left. Institutional memory and luck. Because $3,500/year is out of reach. Tweet 3: I built an alternative in one evening. The 990 pipeline scans public IRS data daily, scores each opportunity against our org profile, and delivers a ranked digest every Monday. First scan found four high-priority funders we'd never encountered. Tweet 4: This is a pattern across the whole sector. Take public data — 990s, census records, foundation annual reports. Wrap it in a dashboard. Charge $1,000 to $50,000 per year. Call it infrastructure. Tweet 5: Before signing any software contract over $100/month, ask two questions: Where does this data actually come from? What would it take to access that source directly? Sometimes the answer is one evening and a Python script. Tweet 6: The 990 pipeline I built for The Matthews is now a Pulley feature. The hard work isn't the packaging. It's solving the problem in the first place. Build the thing that solves your actual problem. Then notice whether other people have the same one. Tweet 7: Most organizations solve operational problems and immediately move on. Treating the solution as overhead instead of asset. The data is public. The problem is probably already solved somewhere in your operations. The question is whether you've noticed.

Tweet 1: Most nonprofits don't lose donors at the ask. They lose them in the gaps. The thank-you that arrived six weeks late. The mid-year touchpoint that never happened. The renewal email sent to someone who already lapsed. The ask is usually fine. The work around the ask is what breaks. Tweet 2: When I mapped out the fundraising operation at The Matthews, I found something uncomfortable. Most of it isn't relationship work. Most of it is administrative processing — logging gifts, triggering acknowledgments, flagging renewals, tracking benefit fulfillment. None of that is the relationship. All of it has to happen for the relationship to survive. Tweet 3: Two types of work. Completely different requirements. Relationship moments: the personal thank-you call, the face-to-face ask, the moment a donor tells you what the organization means to them. These cannot be automated. The moment you try, the donor notices — and it's worse than silence. Administrative processing: everything else. Tweet 4: The failure mode in small nonprofits is that both types of work live in the same person's head, undifferentiated. When that person gets pulled into a crisis — and in a three-person shop, the crisis is always arriving — the processing stops. The gaps open. The donor doesn't notice immediately. But in six months, the relationship has cooled and they can't say why. Tweet 5: Most CRM-driven fundraising gives you a firehose. Here's everything that's due. You sort through it. The cognitive overhead of the sort becomes its own bottleneck. You end up triaging the triage. What you actually need: three to five specific actions, scheduled to specific days, with context attached. Not a firehose. A calendar. Tweet 6: This design principle — separating relationship moments from administrative processing, then systematizing the processing — is valuable before you touch a single piece of software. If you can't describe the processing steps clearly enough to write them down, you don't have a technology problem. You have a clarity problem. And technology will make it worse. Tweet 7: Here's the exercise worth doing before your next major gifts push: List the last three donors who lapsed — not the ones with a reason to stop, the ones who just quietly stopped. What happened in the gaps? Usually nothing. The processing slipped. The relationship cooled. Nobody noticed until the renewal didn't come back. That's an administrative failure. Not a fundraising failure. And administrative failures are fixable.

Tweet 1: Candid charges $3,500 a year for access to public IRS data. Every nonprofit 990 is public record. Free. Queryable through a public API. They wrapped a search box around it. Tweet 2: Small nonprofits run grant development on board connections, Google searches, and whatever the previous ED knew before they left. Institutional memory and luck. Because $3,500/year is out of reach. Tweet 3: I built an alternative in one evening. The 990 pipeline scans public IRS data daily, scores each opportunity against our org profile, and delivers a ranked digest every Monday. First scan found four high-priority funders we'd never encountered. Tweet 4: This is a pattern across the whole sector. Take public data — 990s, census records, foundation annual reports. Wrap it in a dashboard. Charge $1,000 to $50,000 per year. Call it infrastructure. Tweet 5: Before signing any software contract over $100/month, ask two questions: Where does this data actually come from? What would it take to access that source directly? Sometimes the answer is one evening and a Python script. Tweet 6: The 990 pipeline I built for The Matthews is now a Pulley feature. The hard work isn't the packaging. It's solving the problem in the first place. Build the thing that solves your actual problem. Then notice whether other people have the same one. Tweet 7: Most organizations solve operational problems and immediately move on. Treating the solution as overhead instead of asset. The data is public. The problem is probably already solved somewhere in your operations. The question is whether you've noticed.

Tweet 1: Most nonprofits don't lose donors at the ask. They lose them in the gaps. The thank-you that arrived six weeks late. The mid-year touchpoint that never happened. The renewal email sent to someone who already lapsed. The ask is usually fine. The work around the ask is what breaks. Tweet 2: When I mapped out the fundraising operation at The Matthews, I found something uncomfortable. Most of it isn't relationship work. Most of it is administrative processing — logging gifts, triggering acknowledgments, flagging renewals, tracking benefit fulfillment. None of that is the relationship. All of it has to happen for the relationship to survive. Tweet 3: Two types of work. Completely different requirements. Relationship moments: the personal thank-you call, the face-to-face ask, the moment a donor tells you what the organization means to them. These cannot be automated. The moment you try, the donor notices — and it's worse than silence. Administrative processing: everything else. Tweet 4: The failure mode in small nonprofits is that both types of work live in the same person's head, undifferentiated. When that person gets pulled into a crisis — and in a three-person shop, the crisis is always arriving — the processing stops. The gaps open. The donor doesn't notice immediately. But in six months, the relationship has cooled and they can't say why. Tweet 5: Most CRM-driven fundraising gives you a firehose. Here's everything that's due. You sort through it. The cognitive overhead of the sort becomes its own bottleneck. You end up triaging the triage. What you actually need: three to five specific actions, scheduled to specific days, with context attached. Not a firehose. A calendar. Tweet 6: This design principle — separating relationship moments from administrative processing, then systematizing the processing — is valuable before you touch a single piece of software. If you can't describe the processing steps clearly enough to write them down, you don't have a technology problem. You have a clarity problem. And technology will make it worse. Tweet 7: Here's the exercise worth doing before your next major gifts push: List the last three donors who lapsed — not the ones with a reason to stop, the ones who just quietly stopped. What happened in the gaps? Usually nothing. The processing slipped. The relationship cooled. Nobody noticed until the renewal didn't come back. That's an administrative failure. Not a fundraising failure. And administrative failures are fixable.

Tweet 1: Candid charges $3,500 a year for access to public IRS data. Every nonprofit 990 is public record. Free. Queryable through a public API. They wrapped a search box around it. Tweet 2: Small nonprofits run grant development on board connections, Google searches, and whatever the previous ED knew before they left. Institutional memory and luck. Because $3,500/year is out of reach. Tweet 3: I built an alternative in one evening. The 990 pipeline scans public IRS data daily, scores each opportunity against our org profile, and delivers a ranked digest every Monday. First scan found four high-priority funders we'd never encountered. Tweet 4: This is a pattern across the whole sector. Take public data — 990s, census records, foundation annual reports. Wrap it in a dashboard. Charge $1,000 to $50,000 per year. Call it infrastructure. Tweet 5: Before signing any software contract over $100/month, ask two questions: Where does this data actually come from? What would it take to access that source directly? Sometimes the answer is one evening and a Python script. Tweet 6: The 990 pipeline I built for The Matthews is now a Pulley feature. The hard work isn't the packaging. It's solving the problem in the first place. Build the thing that solves your actual problem. Then notice whether other people have the same one. Tweet 7: Most organizations solve operational problems and immediately move on. Treating the solution as overhead instead of asset. The data is public. The problem is probably already solved somewhere in your operations. The question is whether you've noticed.

Tweet 1: Most nonprofits don't lose donors at the ask. They lose them in the gaps. The thank-you that arrived six weeks late. The mid-year touchpoint that never happened. The renewal email sent to someone who already lapsed. The ask is usually fine. The work around the ask is what breaks. Tweet 2: When I mapped out the fundraising operation at The Matthews, I found something uncomfortable. Most of it isn't relationship work. Most of it is administrative processing — logging gifts, triggering acknowledgments, flagging renewals, tracking benefit fulfillment. None of that is the relationship. All of it has to happen for the relationship to survive. Tweet 3: Two types of work. Completely different requirements. Relationship moments: the personal thank-you call, the face-to-face ask, the moment a donor tells you what the organization means to them. These cannot be automated. The moment you try, the donor notices — and it's worse than silence. Administrative processing: everything else. Tweet 4: The failure mode in small nonprofits is that both types of work live in the same person's head, undifferentiated. When that person gets pulled into a crisis — and in a three-person shop, the crisis is always arriving — the processing stops. The gaps open. The donor doesn't notice immediately. But in six months, the relationship has cooled and they can't say why. Tweet 5: Most CRM-driven fundraising gives you a firehose. Here's everything that's due. You sort through it. The cognitive overhead of the sort becomes its own bottleneck. You end up triaging the triage. What you actually need: three to five specific actions, scheduled to specific days, with context attached. Not a firehose. A calendar. Tweet 6: This design principle — separating relationship moments from administrative processing, then systematizing the processing — is valuable before you touch a single piece of software. If you can't describe the processing steps clearly enough to write them down, you don't have a technology problem. You have a clarity problem. And technology will make it worse. Tweet 7: Here's the exercise worth doing before your next major gifts push: List the last three donors who lapsed — not the ones with a reason to stop, the ones who just quietly stopped. What happened in the gaps? Usually nothing. The processing slipped. The relationship cooled. Nobody noticed until the renewal didn't come back. That's an administrative failure. Not a fundraising failure. And administrative failures are fixable.

Tweet 1: Candid charges $3,500 a year for access to public IRS data. Every nonprofit 990 is public record. Free. Queryable through a public API. They wrapped a search box around it. Tweet 2: Small nonprofits run grant development on board connections, Google searches, and whatever the previous ED knew before they left. Institutional memory and luck. Because $3,500/year is out of reach. Tweet 3: I built an alternative in one evening. The 990 pipeline scans public IRS data daily, scores each opportunity against our org profile, and delivers a ranked digest every Monday. First scan found four high-priority funders we'd never encountered. Tweet 4: This is a pattern across the whole sector. Take public data — 990s, census records, foundation annual reports. Wrap it in a dashboard. Charge $1,000 to $50,000 per year. Call it infrastructure. Tweet 5: Before signing any software contract over $100/month, ask two questions: Where does this data actually come from? What would it take to access that source directly? Sometimes the answer is one evening and a Python script. Tweet 6: The 990 pipeline I built for The Matthews is now a Pulley feature. The hard work isn't the packaging. It's solving the problem in the first place. Build the thing that solves your actual problem. Then notice whether other people have the same one. Tweet 7: Most organizations solve operational problems and immediately move on. Treating the solution as overhead instead of asset. The data is public. The problem is probably already solved somewhere in your operations. The question is whether you've noticed.

Tweet 1: Most nonprofits don't lose donors at the ask. They lose them in the gaps. The thank-you that arrived six weeks late. The mid-year touchpoint that never happened. The renewal email sent to someone who already lapsed. The ask is usually fine. The work around the ask is what breaks. Tweet 2: When I mapped out the fundraising operation at The Matthews, I found something uncomfortable. Most of it isn't relationship work. Most of it is administrative processing — logging gifts, triggering acknowledgments, flagging renewals, tracking benefit fulfillment. None of that is the relationship. All of it has to happen for the relationship to survive. Tweet 3: Two types of work. Completely different requirements. Relationship moments: the personal thank-you call, the face-to-face ask, the moment a donor tells you what the organization means to them. These cannot be automated. The moment you try, the donor notices — and it's worse than silence. Administrative processing: everything else. Tweet 4: The failure mode in small nonprofits is that both types of work live in the same person's head, undifferentiated. When that person gets pulled into a crisis — and in a three-person shop, the crisis is always arriving — the processing stops. The gaps open. The donor doesn't notice immediately. But in six months, the relationship has cooled and they can't say why. Tweet 5: Most CRM-driven fundraising gives you a firehose. Here's everything that's due. You sort through it. The cognitive overhead of the sort becomes its own bottleneck. You end up triaging the triage. What you actually need: three to five specific actions, scheduled to specific days, with context attached. Not a firehose. A calendar. Tweet 6: This design principle — separating relationship moments from administrative processing, then systematizing the processing — is valuable before you touch a single piece of software. If you can't describe the processing steps clearly enough to write them down, you don't have a technology problem. You have a clarity problem. And technology will make it worse. Tweet 7: Here's the exercise worth doing before your next major gifts push: List the last three donors who lapsed — not the ones with a reason to stop, the ones who just quietly stopped. What happened in the gaps? Usually nothing. The processing slipped. The relationship cooled. Nobody noticed until the renewal didn't come back. That's an administrative failure. Not a fundraising failure. And administrative failures are fixable.

Tweet 1: Candid charges $3,500 a year for access to public IRS data. Every nonprofit 990 is public record. Free. Queryable through a public API. They wrapped a search box around it. Tweet 2: Small nonprofits run grant development on board connections, Google searches, and whatever the previous ED knew before they left. Institutional memory and luck. Because $3,500/year is out of reach. Tweet 3: I built an alternative in one evening. The 990 pipeline scans public IRS data daily, scores each opportunity against our org profile, and delivers a ranked digest every Monday. First scan found four high-priority funders we'd never encountered. Tweet 4: This is a pattern across the whole sector. Take public data — 990s, census records, foundation annual reports. Wrap it in a dashboard. Charge $1,000 to $50,000 per year. Call it infrastructure. Tweet 5: Before signing any software contract over $100/month, ask two questions: Where does this data actually come from? What would it take to access that source directly? Sometimes the answer is one evening and a Python script. Tweet 6: The 990 pipeline I built for The Matthews is now a Pulley feature. The hard work isn't the packaging. It's solving the problem in the first place. Build the thing that solves your actual problem. Then notice whether other people have the same one. Tweet 7: Most organizations solve operational problems and immediately move on. Treating the solution as overhead instead of asset. The data is public. The problem is probably already solved somewhere in your operations. The question is whether you've noticed.