lesson
30 min

Enrichment: Filling In the Facts

How waterfall enrichment fills in missing facts from several sources, and why you verify every result.

What you will be able to do

Explain waterfall enrichment and run a small version of it in Python.

What enrichment is

A signal tells you that something happened. It rarely tells you everything you need. Enrichment means adding the missing facts to each row: the organization's size, its website, the right department, the job title of the person who decides.

The waterfall

No single source has every fact. So the Temerarii method asks several sources in order and stops at the first good answer. This is called a waterfall, because the question falls from one source to the next until one answers.

  1. Ask source A. If it has a valid answer, keep it and stop.
  2. If not, ask source B.
  3. If not, ask source C.
  4. Verify whatever you got. If nothing is valid, mark the row "unknown". Do not guess.

The last step matters most. A wrong fact is worse than a blank, because a wrong fact ends up in a message.

Run a waterfall

In the code below, each source is a dictionary of made-up organizations. The function checks the sources in order and returns the first answer that passes a simple check: an enrollment number above zero.

If you are on phone data, note that the first run downloads the Python runtime (several MB). You can read the code here first:

source_a = {"Lakeview Middle School": 640}
source_b = {"Northside Library": 0, "Lakeview Middle School": 600}
source_c = {"Northside Library": 310}

sources = [("A", source_a), ("B", source_b), ("C", source_c)]

def is_valid(value):
    return value is not None and value > 0

def enrich(org):
    for name, source in sources:
        value = source.get(org)
        if is_valid(value):
            return value, name
    return None, None

for org in ["Lakeview Middle School", "Northside Library", "Riverbend Youth Center"]:
    value, found_in = enrich(org)
    if value is None:
        print(org + ": unknown (a person should check)")
    else:
        print(org + ": " + str(value) + " (source " + found_in + ")")
Run the waterfall. Then add an enrollment number for Riverbend Youth Center to source_c and run it again.
Python 3
Ctrl+Enter to run | Tab to indent | Python runs in your browser via Pyodide

Notice "Riverbend Youth Center" comes back as unknown. That is the honest answer. A person can look it up by hand, or the row can wait.

Enriching is not messaging

Here is a rule from the Temerarii method that saves organizations from disaster: enriching everyone and messaging everyone are different decisions. You may fill in facts for 500 rows so you can score them. That does not mean all 500 get a message. Most should not. Mixing up those two decisions is how a sending domain gets burned.

When not to enrich

  • Do not enrich personal details you do not need. If a job title is enough, do not collect a personal phone number.
  • Do not enrich from sources you have not approved at the signal gate.
  • Do not enrich people who have already opted out. Their "no" covers this too.

Quiz: Check your understanding

  1. 01In a waterfall, what happens when source A has a valid answer?

    Choose one.

  2. 02No source has a valid answer for a row. What should the machine do?

    Choose one.

Go deeper: Stop Copy-Pasting: The No-BS Buyer's Guide to Zapier, Make, and n8n

Ask. Answer. Build together.