Label gaps, don’t guess
A blank channel field becomes unattributed, not a best-guess “Facebook.” You see the share of revenue we can’t place, which is often the most useful number on the page.
About · Central Islip, NY
Sennum.
A two-desk reporting studio on Islip Ave. We take the exports your store and ad accounts spit out and turn them into numbers you can defend.
Every number we hand over has a formula and a source field behind it.
Sennum started from a recurring problem: a retailer’s Shopify revenue, Meta’s reported purchases and the email platform’s attributed sales added up to more than the bank deposited. Nobody was lying. The files simply used different date formats, SKU spellings and attribution windows, and nobody had written the rules down.
We write the rules down. Then we apply them the same way every month, so a change in a KPI means the business changed, not the spreadsheet.
How the work got shaped
Four rules we added after the data bit us
Each habit below exists because a project once went sideways without it. They are now part of every scope.
- RULE 01 · DUPLICATE KEYS
Count the order once
The first scrub we ran collapsed one order that appeared three times across three exports. Revenue dropped 11% on paper and finally matched the deposits. Now every intake begins with a duplicate-key check on order ID.
- RULE 02 · SOURCE OF TRUTH
Store revenue wins
Ad platforms report what they think they caused. The store reports what was paid. We reconcile to the store and show the gap openly, with attribution windows noted beside it.
- RULE 03 · WRITTEN FORMULAS
No tile without a definition
“Repeat rate” meant two different things in one client’s two spreadsheets. Now each KPI carries a plain-language formula and a named source field before it is drawn.
- RULE 04 · LEAVE CLEAN
You can walk away with everything
Dashboards sit in your own account, with the data dictionary and cleaned tables handed over. If you stop working with us, the reports keep running.
Working principles
What we do when the data is inconvenient
Rerunnable beats clever
Every cleaning step is logged as a rule: date to ISO 8601, SKU zero-padded to four, amount numeric at two decimals. Next month’s file goes through the same rules and any new defect is flagged, not silently patched.
Say what we can’t see
Meta and GA4 will not agree, and we say so. The definitions tab explains the 7-day click / 1-day view window against last non-direct click, so nobody argues over a number that was never comparable.
Ask for the least access
Exports first, view-only GA4 where needed. We never ask for admin credentials or passwords. Working copies are deleted 30 days after hand-off, or sooner on request.
Studio spec sheet
The shop, as a data dictionary
Because you’ll want to know who is reading your files and where they sit.
- location863 Islip Ave, Central Islip, NY 11722
- hoursMon–Fri 9:30am–5:30pm · Sat 10am–4pm · Sun closed
- works inExcel, Google Sheets, Looker Studio, CSV pipelines
- reads fromShopify and POS exports, Meta Ads, Google Ads, GA4, email platforms
- billingFixed fee per project · half at kickoff, half at hand-off · net 14
- quote time48 hours from sample export
- revisions30 days after hand-off
Fit check
Who we’re built for, and who should look elsewhere
A good fit
Small teams with three or fewer data sources
- A store or POS plus Meta Ads and Google Ads
- Under 250,000 rows per month
- Someone who wants one page of KPIs, not forty
- Owners who want the files in their own account
Probably not
Warehouse builds and live streaming data
- Real-time pipelines or custom data warehouses
- Predictive models or machine-learning work
- Multi-store rollups beyond our posted scope, though these we will quote after seeing a sample
Next step
Hand us the worst file you have
Send roughly 50 rows from the export that never reconciles. A fixed-fee scope comes back within 48 hours, in writing.
Send a sample export