Your Booking Export Has 1,600 Rows. Your Boss Wants One Answer. (Code Included)

A long booking export lands in a marketing meeting. Sources, dates, dollars, hundreds of rows. Everyone can see the data. Nobody can tell which source deserves a closer look.

One row per reservation is fine for operations. It’s not how a marketing manager decides where to spend attention.

The Marie Kondo Blueprint is a simple process: group the pile, add a useful measure, and put the most important questions at the top.

Three moves from export to decision

1. Pick a category a manager recognizes. Booking source works because marketing can act on it. A technical ID would look neat and help nobody.
2. Add a business measure beside the count. `COUNT(*)` shows activity. `SUM(total_amount)` shows the booked value behind that activity.
3. Sort for the conversation you need. Ordering by booked value puts the largest financial questions first. Ordering by booking count supports a different conversation.

I ran this against the live Summit Adventures database, the fake adventure tourism company I created to help people learn business analytics, to answer one question: which booking sources produce the most booked value?

SELECT
    booking_source,
    COUNT(*) AS bookings,
    SUM(total_amount) AS booked_value,
    SUM(discount_amount) AS discounts_given
FROM bookings
GROUP BY booking_source
ORDER BY booked_value DESC NULLS LAST;

Results:

| Booking source | Bookings | Booked value | Discounts given |
| — | —: | —: | —: |
| Phone | 312 | $802,046.04 | $25,064.88 |
| Travel agent | 334 | $795,272.83 | $21,432.83 |
| Website | 338 | $785,467.10 | $17,998.37 |
| Referral | 317 | $713,570.24 | $13,738.43 |
| Corporate | 299 | $690,194.48 | $18,976.62 |

The five sources are closer than a raw booking list suggests. A few things stand out:

Phone punches above its weight. Fewer bookings than the website. More booked value.
Referral runs the leanest. Smallest discount total of any channel.
Corporate is small and mighty. Smallest booking count. Still a substantial revenue slice.

This doesn’t tell Summit to abandon any source. It gives the team a cleaner agenda:

1. What’s different about phone bookings that makes their booked value higher?
2. Is the website attracting smaller bookings, or a different trip mix?
3. Are travel-agent discounts negotiated partnership terms, or a pattern worth reviewing?

Sorting isn’t judging. The query is organizing the evidence so the team can ask a better question.

Before you call a source a winner

Booked value isn’t profit. A source can generate a large total and still be expensive to serve.
A high total can hide two stories. Phone leads in value. The website has more bookings. Average value and trip mix would explain the gap.
Discounts need a definition. A travel-agent discount may be a partnership term, not a problem to eliminate.

Group, measure, then sort

When you get a sprawling export, start with one category and one business measure.

Product team: product category and revenue.
Support team: request type and resolution time.
Job search: application source and interviews.

Then order the result to match the decision. Here, `ORDER BY booked_value DESC` puts the largest financial questions at the top. Don’t try to organize every field at once. One useful view beats a perfect dashboard nobody can explain.

This Week’s Action Item

Take one table with repeated records and group it by the category that would matter most to a manager. Add `COUNT(*)` and one value measure, then sort the result by that value measure.