Working an Insight

A cohort-level insight card tells you that something is true about a group of your donors. Clicking the card opens the insight detail page, which shows who is in that group, how Dataro identified them, and what you can change before you act. Everything you need to work the cohort sits on that one page: the donor table, the filter controls, and the export and publish actions.

The Insight Hero

At the top of the page the insight is stated in one line. This is the same sentence that appeared on the card, repeated so that the insight stays visible while you work through the donor table.

How we detected this

Directly below the hero is a one-line explanation of the model behind the insight, with a link to that model's performance page. That page shows how accurate the model is on your own data, and we recommend opening it the first time you use a new insight type.

The pre-filtered table

The table below is filtered to the insight's business logic. You are not building a segment from scratch, you are looking at one Dataro has already built, and the row count is the size of the cohort the insight called out.

Default columns match the ones set for that specific model, so a recurring giving insight shows recurring giving fields rather than a generic donor list.

The Actions Bar

Export, save, and publish live here. CSV and PDF export are always available. Publishing to your CRM, or pushing to Google or Meta, requires you to save the audience first.

You can add filters, remove the filters Dataro applied, and adjust which columns are shown. The table updates as you go, so the row count always reflects what you are looking at.

When you remove a filter Dataro applied, the table is no longer strictly the insight's cohort. That is often what you want, for instance to exclude a region or a recently contacted group, but it does mean you have moved away from Dataro's recommendation. Reset to Dataro recommended returns the table to the original insight cohort in one click, with Dataro's filters and columns restored. Because the reset is always available, you can filter aggressively to understand a group and still return to the starting point.

To adjust the cohort on an insight detail page:

1. Open the insight card from your homepage.
2. Review the table as Dataro filtered it, and note the row count.
3. Add a filter, or remove one of Dataro's, using the controls above the table.
4. Adjust columns if you need a field that is not shown.
5. If the result no longer looks right, click Reset to Dataro recommended.
6. Export, or click Save Audience if you intend to publish.

Refine with chat is a beta control that lets you describe the change you want in plain English rather than assembling filter chips by hand. Type an instruction such as excluding donors contacted in the last thirty days, or restricting the cohort to one state, and the chat converts it into filter chips on the table.

Filter chips are precise, but building a set of them takes time. Expressing something like "recurring givers in Queensland who have not been contacted since June, excluding anyone already in the appeal" takes several passes through filter menus, and each pass is a chance to pick the wrong field. Describing the same thing in a sentence is faster, and it is closer to how you would brief a colleague.

To refine an audience with chat:

1. Click Refine with chat on the insight detail page.
2. Describe the change you want in one instruction. Be specific about fields and values where you can.
3. Read the filter chips it produces above the table.
4. Adjust or remove any chip by hand if it is not what you meant.
5. If the result is wrong enough that fixing it is fiddly, click Reset to Dataro recommended and start again.

Checking what chat produces

We recommend treating the output as a first draft. The chips are inspectable, so you can see which field, operator, and value the chat chose, and compare that against what you asked for.

What beta means here

The chat assistant is still improving, so its accuracy at translating instructions into filters will continue to get better, and larger updates to the chat experience are planned.

Export the cohort as CSV or PDF at any point. No save is required, and the export reflects the table as it currently stands, including any filters you added and columns you changed. CSV is the working file, which you can hand to a mail house, load into a spreadsheet, or reconcile against your own records. PDF is the shareable version, useful when someone needs to see the cohort without being able to edit it.

Publishing is different, because it writes data into another system. These three destinations all require a saved audience:

Until you save, those controls show a Save first indicator rather than failing without explanation, so you can see what is missing before you try to publish.

Saving an audience

To save an audience from an insight detail page:

1. Click Save Audience.
2. Enter a name you will recognize later. We recommend including the insight and the month if you run the same campaign repeatedly.
3. Add a description covering why you chose this cohort, what you excluded, and which campaign it feeds.
4. Confirm the save. The filters and columns you selected are attached to the saved audience.
5. The publish controls become available, and the Save first indicator clears.

The name tells you which audience this is, and the description tells the next person why it exists. We recommend filling in both, because the description is the part that is easy to skip and hard to reconstruct later. A description that reads "RG churn insight, October appeal, excluded anyone contacted in September" answers the questions a colleague would otherwise bring to you, and it answers them for you when you repeat the same campaign next quarter.

Where the audience goes next

An audience saved from an insight lands in the Dataro audiences tab, not Saved, tagged as having come from an insight. It is an audience in its own right rather than a snapshot of this page, so it keeps that tag permanently and it can be published later rather than immediately. Saving and publishing are separate steps, so save the audience while the cohort is in front of you and publish when the campaign is ready to go out. Once published to your CRM, the CRM Queue tab shows you what happened to the push, so the trail from insight, to saved audience, to CRM is continuous.

Why can I export but not publish?

Because export only reads the cohort, while publishing writes to another system. CSV and PDF are available at any time, with filters and columns as they currently appear on screen. Pushing an audience into your CRM or an ad platform creates records other people rely on, so Dataro requires a named, described, and saved audience first.

What does the "Save first" indicator mean?

That the audience has not been saved yet, so the publish path is not available. It is a prompt, not an error. Click Save Audience, and the publish controls become active. Whatever filters and columns you adjusted are attached to the audience when you save, which is what makes it reproducible rather than a one-off list.

What if I filter too far and lose the audience?

Click Reset to Dataro recommended. One click restores Dataro's original filters and columns, and nothing you filtered is permanent until you save.

Can chat do anything the filter controls cannot?

No. It produces the same filters you could add yourself, so it cannot go beyond what the filter controls already offer.

The same donor is in two insights. Did Dataro make a mistake?

No. The five models have different eligibility rules, several of them mutually exclusive, so overlap is rare, and when a donor does qualify for two the donor is genuinely relevant to both actions. Individual donor alerts are limited to one per donor per day, so you will not get two alert cards for the same person on the same day. We recommend sequencing the contact rather than sending both at once: anything time-limited first, a stewardship moment before an ask, and your own time on the larger relationship rather than on the mail appeal that will reach the donor anyway.

Can I choose which insight types my organization sees?

Not today. Dataro decides which insight types exist and which thresholds they fire on, and everyone in your organization sees the same cards.

Can I mark an insight as acted on?

No. Insights have no acted-on state, so a card looks the same after you have worked it and there is no done, dismissed, or snooze control. Use the audience trail instead: an audience you save from an insight lands in Dataro audiences tagged to that insight, and the CRM Queue tab leaves a second, timestamped record of anything you published.