NationBuilder Insights on Demand: Data Appendix

Use this page to look up any custom field or tag that the Insights On Demand models write to your NationBuilder records. For an overview of what the service is, what data it reads, and how a run works, see the program overview. Tag and custom field names are exactly as they appear in NationBuilder, so you can search for them directly.

Each model has its own section, explaining its custom fields first and its tags after.

Model 1: Donor Lifetime Value

Model 2: First Donation Prospects

Model 3: Email List Health

This model looks at each donor in your nation and produces two estimates. The first is how much that donor is worth over a 36-month span, and the second is a score for how likely they are to give again in the next 12 months.

The model stores these results in your nation as four custom fields and two tags.

Who gets it: Donors in the Major donor segment.

These are donors whose single donations reach the major-donor line. The suggested action is to cultivate them, which means building a closer relationship with them over time.

Who gets it: Donors who aren't major donors, are in the Highest or High band, have a chance under 30% of giving again in the next 12 months, and made a one-time donation in the last 24 months.

These are valuable donors who look less likely to give again. The suggested action is to re-engage them, which means reconnecting with them before they drift further.

This model shows how valuable each donor is likely to be and how likely they are to give again, and it flags your major donors. It needs a record of donations with dates. Anything dated before 1990 is ignored as irrelevant data. Dollar estimates are only added when the model's guesses land within 50% of what your donors actually gave in the last 12 months. If they don't, those stay blank and everything else is still added.

Major donors do not get a total dollar estimate, and their chance of giving again is less certain because there's often less history to learn from and their giving amounts and frequency can vary significantly.

This model looks at everyone in your nation who hasn’t donated yet and ranks them by how likely each person is to make a first donation in the next 90 days. Because the ranking compares people only with one another, it tells you who is more likely to give than whom, but it doesn’t predict how many people will actually give.

The model learns those patterns from your nation’s own donation history, so it needs at least 200 first donations on record before it will add any scores.

Who gets it: People in the High likelihood group.

These are people who haven't donated yet and are most likely to make a first donation. The suggested action is to ask them to make a first donation.

📌 Note — The tag name keeps the word "gift" from the model's original name. It means the same as "first donation prospect."

This model needs at least 200 people who made a first donation in roughly the last three years, and at least 20 of those in the last 90 days.

It also tests itself on those recent 90 days, and it only goes ahead if it predicts better than just picking whoever was active lately. If your donation history doesn't meet all of that, nothing is added, so you never get a ranking built on too little information.

This model places each person in at most one segment, based on how they engage with your emails. It also ranks the people who click the most. It writes no custom fields.

A person gets a segment only if they can still be emailed (opted in, not marked do-not-contact, and with a usable address) and have received at least one of your campaigns without a bounce. Repeat bouncers are the exception.

Every tag below describes how someone is engaging with your emails right now. Every suggestion comes back to one idea: attention is valuable, and it's easy to waste. When people are reaching for more, invite them in. When they're already with you, protect their attention. When they've drifted, give them room to come back without crowding them.

A regular is someone who clicked in at least three different calendar months. Each tag below explains what it means, what action we suggest, and why.

Who gets it: People who can still be emailed, whose last two campaign emails both bounced within the last 180 days.

Your emails to them keep bouncing, so they may not be getting your emails at all. Mailbox providers count repeated bounces against the sender, which hurts your deliverability. That affects how well your emails reach everyone on the list. The suggested action is to check the address for common typos and fix it if you can. If you can't fix it, stop sending to that address, and reach out to the person to confirm the best contact information for them.

Who gets it: People who are clicking a noticeably bigger share of your recent emails than they did in the 60 days before.

They’re reading more of your emails lately, and their interest is up. The suggested action is to ask them to get more involved, such as taking an online action, sharing content with a friend, or coming to an event. Make one clear, well-timed ask of them rather than adding more email.

Who gets it: People who clicked in at least three different calendar months, and clicked in the last 60 days.

These are regular readers who keep clicking your emails, so they’re your most active base to work from. Their attention is valuable, and it’s easy to waste. The suggested action is to steward them, which means asking what they care about and what they want to see more of, then giving them more of that. Be considerate about what you send and how often, so you don’t wear out their welcome.
​
​Bonus: These folks make great "control" audiences for A/B testing of content or subject lines.

Who gets it: People who clicked in at least three different calendar months, but not in the last 60 days.

These people used to read your emails regularly, and they’ve stopped clicking lately. They’re good candidates for a win-back campaign. A regular, business-as-usual or newsletter email is unlikely to bring them back, so the suggested action is to win them back with a different message, tone, or channel than the emails you usually send them.

Who gets it: People who clicked for the first time in the last 60 days.

This model looks at everyone in your nation who hasn’t donated yet and ranks them by how likely each person is to make a first donation in the next 90 days. Because the ranking compares people only with one another, it tells you who is more likely to give than whom, but it doesn’t predict how many people will actually give.

Who gets it: People who last clicked 60 to 120 days ago, or came back after a long gap, and who aren't regulars.

These people have gone quiet for a while. The suggested action is to re-engage them gently. Keep the emails light, don't pile on, and give them room to come back when they're ready. If you already know someone in this group has acted with your organization before, such as by donating or volunteering, you may want to reach out to them more directly and listen to what they have to say. This model only sees email activity, so it can't tell you that.

Who gets it: People who haven't clicked in the last 120 days, and who aren't regulars.

These people still get your emails, but they haven't clicked in a long while. The suggested action is to trial sending them fewer emails. Another option is a sunset campaign, which is a final message asking whether they still want your emails. Don't make fewer emails a standing rule for everyone until you've seen how it goes.

📌 Note — Sending less email to people who aren't engaging can help your overall deliverability and inbox placement.

The top 25 people who clicked the most campaigns in the last 30 days each get one rank tag. These tags don't come with a suggested action, and a person can hold a segment tag and a rank tag at the same time. These show the most active clickers on your list at the moment the model runs.

Who gets it: Clickers ranked 1-3.

Who gets it: Clickers ranked 4-5

Who gets it: Clickers ranked 6-10

Who gets it: Clickers ranked 11-25

This model needs emails sent from NationBuilder, with their clicks and bounces. There's no overall minimum of activity, so it always runs. However, a person you can email gets a label only if they were sent at least 2 emails in each of the last two 60-day stretches.

Anyone with less history gets no label, though repeat bouncers and top clickers are still marked. "Rising readers" also requires enough past sending for the model to learn from, and without it, nobody is marked rising.

Every model that writes values also writes to this field. Check it to see which version of each model produced a value, and to compare when the model last ran.