Before any analysis can exist, the foundation has to be real.
explōrātiō takes your files as they are today — platform exports, spreadsheets that grew at the edges, systems that do not talk to each other. You get back a single base with written rules and a dictionary, ready for your team to use on Monday — without spending all of Friday on “cleaning the table for the meeting.”
Data Clean
Input → processing → output
What goes in, what runs in the middle, what you take away — and the project timeline below.
You send files in whatever format you have today. We agree in writing what needs to happen — dedupe rows, standardize campaign names, merge two sources, calculate margin, and so on — and return a ready base with documented rules, so your team’s next step does not depend on guessing what was done.
Raw data
Clean data
Wiki
002Who it's for
For anyone tired of spending Friday fixing spreadsheets for Monday’s meeting.
From marketing to data, everyone loses hours stitching files together, renaming columns, merging platform exports. Here that work becomes a one-time delivery — with written rules and a base ready for the next use, without becoming the fourth parallel spreadsheet nobody knows is official.
Media and marketing teams
Who unifies platform exports
Each platform, each client, each publisher delivers reports in a different format. Here everything becomes a single model, with taxonomy and field names documented — ready to feed the dashboard, MMM, or client report without the intermediate step of “let me just standardize first.”
Analytics teams
Who needs clean input for the model
Models, dashboards, and pipelines need clean input. You get a field dictionary, written transformation rules, and an agreed output format — so your team stops carrying business rules only in the head of the analyst on vacation.
Data science teams
Who needs a documented base for the next model
Value recalculation, harmonization across sources, and adjustments that support the next model — with process documentation so the next round does not start from zero. Delivered the way your team integrates: CSV, SQL, Parquet, whatever is in scope.
003What lands on your desk
Four things that always arrive, in any package.
Whether the package is Light or Enterprise, delivery has four fixed elements. What changes between packages is size — number of sources, rows, and rules — not the shape of the delivery.
Single base, in the format your team uses
Final consolidated file in formats agreed in scope — CSV, XLSX, SQL, Parquet, or other. Ready to open in Excel, import into BI, or ingest into the pipeline, without the intermediate step of “I need to convert first.”
Field dictionary
Each column in the base has: technical name, business-language description, data type, fill rule, and examples. No more “what does this field mean again?” three months after delivery.
Documented transformation rules
Everything done to the base — what was removed, merged, recalculated, renamed — is listed and explained. Auditable by whoever arrived later, repeatable by your team if the next load comes in the same format.
Single adjustment window
After delivery, 15 to 30 days (depends on package) for one consolidated round of adjustments. Anything after that, or in a second round, is treated as new scope — no gray zones, no “just one more thing” becoming habit.
004Our approach
You are not buying hours of data work. You are buying delivery.
Billing by hours worked hides what was done and opens the door to endless project growth. Here the model is different: scope fixed in writing, price fixed before starting, delivery defined before execution. You know exactly what you will receive, how long it will take, and how much it costs — before the first line of code runs.
Scope fixed in the briefing document, before execution
Fixed price per package, no surprise overtime at the end
Delivery defined in agreed files, model, and documentation
Single adjustment window after delivery, in one consolidated round
005Five situations Data Clean already solves
Five situations Data Clean already solves.
Each scenario describes a concrete situation — media team with five platforms, dashboard that breaks on every load, SMB trying to close the month — what was happening before and how the package arrives to solve it. Illustrative scenarios, based on typical market cases.
— Case 01Media team · performance
Five platforms, three taxonomies — and the 2 p.m. meeting with no numbers.
Before
Meta, Google, TikTok, LinkedIn, and local publisher in different formats
Same event with different field names
Mixed currency with no flag
After
Single weekly table (CSV + XLSX)
Harmonized taxonomy by channel
Ready for Looker
Outcome Three business days per week recovered — meeting on time.
Pro package3 sourcesCSV + XLSX
— Case 02BI team · analytics
The dashboard breaks on every load — and the field “changes meaning” every week.
Before
Fields with no written definition
Duplicate keys across sources
Business rule only in the analyst’s head
After
Dictionary and README per base
Unique keys validated before sign-off
SQL output ready for ingestion
Outcome The team goes back to modeling — not copy-paste-and-rename.
Business package7 sources · up to 1M rowsSQL + CSV
— Case 03SMB · e-commerce
“Which SKU sold most?” — every Friday the answer is late.
Before
Order, ad, and cost in separate files
SKU with no master code
Real margin only after month close
After
Order × channel × SKU consolidated
Single documented margin rule
Weekly XLSX for the founder
Outcome SKU and channel decision without hiring a data team.
Light package1 source · up to 20 fieldsXLSX
— Case 04Agency · offline media · MMM
TV, cinema, or OOH — weekly impacts and investment for the mix.
Before
Each vendor calculated impacts differently
Unexpected granularities requiring revision
Monthly data forced into weekly series
After
Values unified by publisher or platform — same logic across the base
Granularity adjusted for modeling, with totals preserved relative to the initial value
Weekly structure aligned for MMM and marketing mix
Outcome Coherent weekly base to run mix models without rework by vendor.
Pro packageTV · cinema · OOHMMM-ready
— Case 05Client · planning · analytics
Non-advertising investments and sales — when the calendar does not align.
Before
Monthly sales in large bases, manually recalculated to weekly
Non-monetary marketing efforts (sales force)
Events and sponsorships with no actionable direct impact
After
Weekly distribution by statistical methods — ready for the model
Non-monetary data treated as exogenous variables
Indirect actions modeled with Adstock and appropriate transformations
Outcome Exogenous variables and sales on the same MMM calendar.
Business packageexogenous · salesAdstock
Illustrative scenarios · based on typical market cases
006Fixed packages
Four sizes. What counts is the volume of your base.
Four packages as one-time delivery. What changes between them is size — how many sources, rows, and transformation rules. In the initial conversation we look at a sample of your base and indicate which package it fits; if it sits in between, we choose the smaller one and treat the excess as an add-on.
Light
R$ 3.947
Sources1 source
Rowsup to 10K rows
Fieldsup to 20 fields
Rulesup to 10 rules
Transformationsup to 10 transformations
Output formats1 format
Documentationcomplete
Pro
R$ 5.947
Sourcesup to 3 sources
Rowsup to 100K rows
Fieldsup to 50 fields
Rulesup to 20 rules
Transformationsup to 10 transformations
Output formats2 formats
Documentationcomplete
Business
R$ 13.947
Sourcesup to 7 sources
Rowsup to 1M rows
Fieldsup to 100 fields
Rulesup to 50 rules
Transformationsup to 20 transformations
Output formats2 formats
Documentationcomplete
Enterprise
R$ 19.947
Sourcesup to 15 sources
Rowsup to 5M rows
Fieldsup to 140 fields
Rulesup to 70 rules
Transformationsup to 50 transformations
Output formats2 formats
Documentationcomplete
For agencies that want to deliver Data Clean under their own brand, and for data teams that want the package in the technical standard they integrate, specific formats by conversation. WhatsApp · reply within 1 business day.
007When scope grows
When scope grows mid-project.
Data work is about scope, and scope sometimes changes — a new source appears, an extra field enters, a calculation rule needs adjusting. Instead of hiding the rule, we put the table in plain sight: what triggers an add-on, what becomes new scope, and what each case costs on top of the contracted package.
Targeted add-on
+35%
Change up to 25% of signed scope
On the initial signed package value. Ex.: Light package (R$ 3.947) → increase of R$ 1.381.
Broad add-on
+50%
Change above 25% of scope
Change above 25% of signed scope. Ex.: Light package (R$ 3.947) → increase of R$ 1.974.
Criterion A
25%
Limit by transformation rules
Ex.: Light package — 10 transformation rules in scope. Request changes 2 rules (20% of scope) → targeted add-on (+35%). From the 3rd rule changed (≥30%) → broad add-on (+50%).
Criterion B
25%
Limit by fields impacted
Ex.: Light package — 20 fields in scope. Request impacts 4 fields (20%) → targeted add-on (+35%). With 5 fields (25%), still targeted; from the 6th field impacted (≥30%) → broad add-on (+50%).
Adjustment window
15–30d
One round after formal delivery
Requests consolidated within the deadline. Written sign-off (email counts). After that = new scope.
Becomes new package
New
Outside add-on — quoted separately
Different base, period, or dictionary; calculated field outside what was agreed; format not foreseen; request after the window.
Criteria A and B measure the % of scope changed · higher index prevails · targeted add-on +35% or broad +50% on the package · all within signed scope
Reference clause in scope: the client has 15 to 30 calendar days after delivery to request adjustments, in a single consolidated round. Requests after that deadline, or in additional rounds, will be treated as new scope and quoted separately.
008 / Next stepInitial conversation · no commitment
Thirty minutes to see if it makes sense. Nothing to buy.
The initial conversation is to look at a sample of your base and say frankly: there is Data Clean here that solves it, which package it fits, or your team can still handle it alone. If yes, we close scope, timeline, and format on the spot. If not, you leave with the diagnosis anyway.
Fixed package, within signed scope. Usable the day after delivery, auditable by whoever arrived later, with written sign-off. No “almost ready,” no endless revision round, no delivering half.
Data
Ready to use
FilesProcessed base in contract formats — CSV, XLSX, SQL, or as agreed
ModelSingle field structure for BI, performance, or ingestion
→
Documentation
No oral memory
DictionaryEach field and fill rule documented
ProcessREADME, flow, and transformations described for the team
→
Audit
Versioned trail
AssumptionsRecorded in writing
VersioningTraceable package
TransformationsExplicit steps and rules
ValidationPre-acceptance checks
→
Close-out
Cycle closed
AdjustmentsOne consolidated round in 15–30 days after formal delivery
Sign-offWritten confirmation; new request becomes new scope