I take a page you can see in your browser — a catalog, a directory, a listing — and hand back
a file you can sort in Excel. Local to Northwest Arkansas, working with teams anywhere.
Twenty rows, same format as the full job, no charge and no card. If the shape is wrong, you have
lost nothing but the email.
oak-side-tables.csvdelivered
ProductPriceChangeCity
Oak side table1000−99BentonvilleOak dining chair24512RogersWalnut shelf1800FayettevilleBirch stool95−5Springdale
1,000 rows · 5 columns0 empty · 0 duplicates
The part most quotes leave out
Anyone can pull text off a page. The work is what happens after: values that look like numbers
but are stored as text, rows that appear twice, cells that are empty. This is one job, shown before
and after.
A supplier price list, three rows of it. The rest of the file behaved the same way.
Product
Price
Change
City
Vendor
Oak side table
$1,0001000
(99)−99
bentonvilleBentonville
ACME LTD Acme Ltd
Oak side table
$1,000
(99)
bentonville
ACME LTDtwo identical rows, one kept
empty row
The full run: 1,000 rows from a 50-page public catalog in 64 seconds, no empty cells, no duplicates.
Every delivery ships with a README sheet naming the source, the date of the pull, and the row count —
so a year from now you can still tell where the numbers came from.
No call, no discovery deck. Three steps, and you see the output before any money moves.
You send a link and the fields you care about.One sentence is enough — "product name, price, stock status, city". If the page is public, that is all I need to start.
You get twenty real rows back, free.Same columns, same cleaning, same README as the full job. It is the actual output, just shorter, so you can judge the format instead of imagining it.
You pick a package, I deliver the rest.CSV plus Excel, README with the source and the pull date, and the before/after for every row I dropped or merged. Single page: back within 24 hours. Anything larger comes with a date before I start.
Four things I do
Scrape a public pagefrom $30
You give a URL and the fields you care about. You get CSV plus Excel, UTF-8 so it
opens clean, with the columns named the way you asked for them.
$30 one page, up to 1,000 rows
$60 up to 5 pages, or one page above 1,000 rows
$120 up to 20 pages
Clean what you already havefrom $25
The export nobody wants to touch. Duplicates merged, numbers turned into numbers, mixed date formats
flagged rather than guessed, and a log of every change so nothing disappears quietly.
$25 one sheet, up to 5,000 rows
$55 up to three sheets, or a messy multi-source merge
$110 a workbook that has to be rebuilt, not repaired
Watch prices on a schedulefrom $35
The same products checked across the stores you compete with, with the change since last time.
Built for people who find out about a price cut three weeks late.
$35 one run, delivered as a sheet with a change column
$85 weekly for a month, four deliveries
$180 daily for a month, plus a note when something moves more than you expect
Summarise and analyse itfrom $40
The file you already have, turned into a page you can act on: pivot tables that recalculate, formulas
instead of pasted numbers, and a short written answer to what actually changed and why —
not a dump of twenty charts.
$40 one sheet, a one-page readout with the three numbers that moved
$75 up to five sheets, with a month-over-month comparison
$150 a monthly report you can forward to someone else, refreshed on a schedule
Something not on this list
A larger set, a recurring pull, a format that isn't CSV, or a page that needs judgement instead of rules —
send the link and one sentence about the output you want.
I quote against the real page count, not a guess, and you get the quote before anything
starts.
Hourly work, when the scope is genuinely unclear: $40/hour, capped at an amount we agree first.
What I turn down
Logins, paywalls, and anything behind an account. I read pages the way a visitor reads them.
If a site's terms forbid automated collection, I say so before you pay rather than after. That has cost me
work; it has not cost anyone a surprise.
Starting
The three fixed prices live on Fiverr, where the order, the files and the money are all protected by the
platform. If you would rather work directly, say so in the form below and we agree the scope in writing first.
Card, Apple Pay and Klarna run through Stripe and work from any country. Cash App is United States only: on
a computer you get a code to scan. Direct transfers carry no platform protection — if you want the order
covered end to end, use the Fiverr button above.
Direct payment (invoice) is available once we have worked together before. First orders go through the platform.
Keep it running, monthly
Some files are not one-off. If the pages you care about change every week, buying a new pull each time is
worse for both of us — the setup cost is the same, the value is in the cadence. Cancel any month; there is
no lock-in and no setup fee.
Keep it current$45 / month
One re-pull a month of the pages already agreed, corrections included, README updated with each pull
date. For a supplier list or a directory you quote from that drifts slowly.
Weekly re-pull, a change report each time — what moved, by how much, which rows
appeared or vanished — and fixes inside 24 hours if a site changes its markup. For competitor prices,
stock status, or anything where a week is too long to be blind.
The data refreshed, plus a one-page summary written for a person: the three numbers
that moved, the likely reason, and what to check next. Sized to the number of sheets after a look at
the file — ask and I quote the exact figure.
Paid by card, Apple Pay, Klarna or Cash App through Stripe; the subscription is charged monthly until
you cancel, and you can cancel from the receipt Stripe emails you.
After the file lands
The part that decides whether a data job was worth it is what happens in the two weeks after delivery.
So it is written down rather than implied:
Seven days of free corrections. If a value is wrong, a row is missing, or a column is not what you asked for, tell me and I fix it — no charge, no argument, inside one working day.
Thirty days of protection against the page changing. If the site changes its structure and the file no longer matches reality, I re-pull it once for free inside that month.
The README stays true. Every delivery records the source URL, the pull date and the row count. On a monthly plan each new pull adds a dated line, so you always know which file came from which week.
Nothing disappears quietly. Rows I drop or merge come with their before/after value. If I am unsure about a date, I flag it instead of guessing.
Cancelling a plan is one click. No notice period, no exit fee, and the last file you paid for stays yours.
README — shipped with the filesheet 1 of 2
Source
oak-side-tables.example/catalog
Pulled
2026-10-27 16:04 CDT
Rows
1,000 (from 50 pages)
Dropped
2 — both listed below
Duplicates
merged, 1 kept
row 41 — Oak side table · $1,000 · (99) · bentonville · ACME LTD
→ kept as Oak side table · 1000 · −99 · Bentonville · Acme Ltd (identical duplicate removed)
row 77 — empty row, no values in any column → removed
Questions I get asked
Is scraping a site legal?
Public pages — the ones a logged-out visitor can read — are the normal case. Where a site's terms forbid automated collection, I say so before you pay, not after. I do not do logins, paywalls, or anything behind an account.
What if the page changes a month later?
The README records the source URL and the pull date, so the file stays auditable. If a site changes structure and you want a fresh pull, the next run is a new job — but on a repeating job I re-check the selectors each time rather than shipping a silent half-empty file.
What do I actually receive?
CSV and Excel, UTF-8, one row per item, columns named the way you asked. Plus the README sheet: source, pull date, row count, and every row dropped or merged with its before/after value.
How do I pay?
First order goes through Fiverr, so the money and the files are both covered by the platform. For repeat work, card, Apple Pay, PayPal or Cash App below — the amount is pre-filled.
Tell me what you need
A link is enough to start. I answer within one working day, and the first twenty rows cost nothing.
Prefer email? hello@chehangdata.com