Lesson 1.2.3
Turn any directory or listing page into a table
What this costs
Charged per row extracted, the same as any other builder.
Before you start
A URL of a page that lists companies.
Most of the markets worth selling into are not on Google Maps. They are on an association member page, a supplier directory, a conference exhibitor list, a government register. Web to Table turns any of those into rows, and this is the move that separates Kuration from a database.
Finding the page worth extracting
The best source pages have three properties: they list many companies at once, membership means something, and no database has bothered to index them. Trade association member lists score on all three. So do accreditation registers, tender portals, chamber of commerce directories, and the supplier lists that large buyers publish.
- A member directory for the trade body your buyers belong to.
- The exhibitor or sponsor list from an event they attend.
- A certification register, where being listed is itself a qualifying signal.
- A "our partners" or "where to buy" page on a company that sells alongside you.
- 1
Start a new research project and choose Web to Table
Paste the URL of the listing page. Use the page that shows the companies, not the site homepage.
- 2
Say what one row is
Describe the unit you want: one row per company, one row per member, one row per exhibitor. This is what stops the extractor returning navigation links and page furniture as rows.
- 3
Name the fields you want off the page
Company name is always one. Then take whatever the page gives you for free: membership category, city, stand number, accreditation level, member since. These become filters later at no extra cost.
- 4
Run it on one page first
Extract a single page and read the result before pointing it at a paginated directory. Ten rows is enough to tell whether the field mapping is right.
- 5
Then extend across pagination
Once one page is clean, run the rest. A 400 member directory extracted correctly is worth more than most bought lists, because nobody else has bothered.
Pointed at the homepage instead of the listing page
- What you see
- A handful of rows that are actually menu items
- The fix
- Navigate to the page a human would scroll to read the list, and use that URL.
Did not describe what a row is
- What you see
- Rows split at the wrong boundary, one company across three rows
- The fix
- State the unit explicitly: one row per exhibiting company.
Ran the whole directory before checking one page
- What you see
- Hundreds of rows with a misaligned column
- The fix
- Always extract one page, read it, then extend.
The listing is behind a login or a search form
- What you see
- Empty or partial extraction
- The fix
- Use the PDF route if there is a downloadable directory, or the Chrome extension if it is a logged in platform.
Check yourself
0 of 2 answered1.Which page makes the best source for Web to Table?
2.Why extract the extra fields a listing page offers, like membership tier or stand number?
Try it yourself
Find the member directory of a trade association your buyers belong to. Extract one page of it, with company name plus at least two fields the page gives you for free.
- You have done it when
- Ten or more clean rows where the extra fields are populated and usable as a filter.
- Credit budget
- One page of extraction.