Methodology last updated 2026-07-28
This is how we measure observed purchases in Panama, invoice by invoice.
A panel is only worth its method. This page documents the source, the consented base, the aggregation rules and the limits of the Cuestio Panama panel, in plain language. If something is not here, ask us.
1The source: the DGI electronic invoice
Panama operates an electronic invoicing system administered by the Dirección General de Ingresos (DGI), Panama's tax authority. Electronic invoices covered by that system carry a QR code and a unique identifier (CUFE) that can be consulted with the DGI. Not every fiscal document issued in Panama is a compatible electronic invoice. That is the first boundary of coverage: the panel measures what its documents allow it to measure.
Users of facturillo scan the QR on their invoices voluntarily. Where the document format allows it, the system extracts the structured information: merchant (tax ID and branch), date and time, items with quantity, paid price, discount and tax. Payment method is included only when it is present in the document and captured correctly. Document validation, data extraction and product classification are three separate steps, each with its own quality controls.
The result is a record of documented purchases, not recollections. No purchase diaries and no shopper recall: in the analytical base, every product line comes from a DGI electronic invoice. Where CUFE lookup is available, the document can be checked against the DGI record.
2The base: consented profiles
Cuestio computes exclusively on data from profiles that switched on explicit consent inside facturillo. As of August 3, 2026:
More than 85% of accepted records are DGI electronic invoices captured via QR. The rest are manual expense entries. Manual entries contribute no product lines. All price, basket and line analytics run exclusively on the DGI electronic invoices. That exclusion sits in the panel's architecture, not in a revisable policy.
The ladder of bases, from the widest to the strictest. The facturillo community exceeds 17,000 registered users, with more than 400,000 valid records captured. Of that community, over 12,000 profiles hold active consent and form the panel, and more than 3,800 were active buyers in the best month of the last six months. Cumulative spend observed in the consented panel exceeds $14 million. That figure applies a per-invoice cap to contain outliers. When we publish a figure, we say which of these bases it comes from. Community figures describe facturillo's reach. Every analysis runs only on the consented panel.
The facturillo app receives on average more than 45,000 submissions per month in 2026, scanned invoices and manual entries combined. The number entering the consented, accepted base is lower, after validation, compatibility and quality checks. For context on method, not size: the region's traditional consumer panels are built on assisted purchase diaries across thousands of households. The Cuestio panel is built on fiscal documents, line by line.
What "consented" actually means
Nobody is in the panel by default. Consent is born switched off and is activated only by joining Puntos Facturillo, in front of a screen that names these uses explicitly. Each acceptance is recorded server-side with its date and its version. Consent can be revoked anytime from the app. Cuestio computes exclusively on that base: the more than 12,000 profiles at the cut.
The full detail: screen, ledger, revocation and legal anchors
Nobody is in the panel by default. Consent is born switched off when the account is created, and switching it on is an explicit decision tied to real value: it happens when a user joins the Puntos Facturillo rewards program. The consent screen says, in its own words, that data will be used to "Generate anonymous market reports" and "Share anonymous trends with partner companies". The same screen states the rights: leave the program anytime from the app settings, your name and email are never shared, partner companies receive only aggregated statistics, and deleting the account deletes the account's data according to the retention and deletion policy.
Every acceptance and every withdrawal is recorded server-side with its timestamp and the exact policy version in force at that moment. The version is never supplied by the client: the server stamps it. The result is an auditable consent ledger, acceptance by acceptance.
Consent can be revoked anytime from the app, by leaving Puntos Facturillo or by deleting the account. Withdrawal stops future processing under that consent. What it does not do is retroactively invalidate aggregates already lawfully published, because they contain no personal data. The consent screen itself says so: what is already anonymized stays anonymous.
The legal anchors, by name. The data controller is facturillo Inc. The facturillo Privacy Policy (updated July 8, 2026), section "Monetización de Datos y Análisis": aggregated, anonymized market statistics, with a separate opt-in tier for pseudonymized profile analyses via Puntos Facturillo. The facturillo Terms (updated July 8, 2026), sections "Venta y Compartición de Datos Anonimizados" and "Consentimiento para la Recopilación y Procesamiento de Datos": any sharing of anonymized usage data is conditioned on the explicit consent granted by joining Puntos Facturillo. Cuestio operates within that condition, the stricter of the two.
A transparency note on the figures: the public numbers on facturillo.com are rounded and cover the whole app, so they are larger than Cuestio's. The difference is the scope. Cuestio publishes only the consented, accepted base: the stricter set, on purpose.
3What we measure
- Paid price per item, with discounts applied where the document records them: the figure a shelf audit normally does not see, because the shelf shows the displayed price, not the final till price.
- Basket composition: More than 1.6 million purchase lines extracted from accepted electronic invoices, linked where possible to a normalized catalog of 54,000+ products with brand, category (25 categories) and barcode.
- Cross-chain behavior: when the same consented profile submits accepted invoices from several merchants, the panel enables cross-shopping, wallet-leakage, brand-switching and penetration metrics, always within the observed chains.
- Geography: More than 14,000 geolocated branches, assignable to province, district and corregimiento where reliable observed locations exist.
- Shopper profile: age, gender and region, in aggregates only and only where the data exists and clears the sample thresholds.
4What clients receive (and what they never receive)
Everything Cuestio publishes or delivers is anonymous and aggregated: shares, medians, trends and patterns that cannot be traced back to any person. No client receives purchase histories, individual profiles or identifiers of any kind. Three rules stand behind every published figure:
- The floor of 30. Every published behavioral figure is built on at least 30 distinct consented profiles, and price statistics add metric-specific thresholds on top. What does not reach the floor does not get published.
- Curated aggregates only. We publish editions and reports, not raw-data access or open queries. In public figures, retail chains additionally appear anonymized. Commissioned private reports may name chains, under the agreed scope and these same thresholds.
- Living consent. The panel is built only on profiles that switched their consent on, and that consent can be withdrawn anytime from the app. Aggregates already published contain nothing personal to withdraw.
The full data-governance documentation, with the technical aggregation, suppression and anonymization rules, is available to clients and auditors on request.
5The limits, stated up front
The panel is opt-in: it is not a probability sample of the Panamanian population. It tends to overrepresent urban, digitally engaged shoppers who keep and submit their invoices. Coverage also varies by merchant, document type and technical compatibility: purchases backed by processable electronic invoices weigh more in the panel than purchases documented in unsupported formats. By design, Cuestio:
- reports panel-observed shares, gaps and trends, not unqualified absolute market sizes. We do not weight or project results to the national population.
- states the base (n) behind every published figure and says whether it counts profiles, records, transactions or lines.
- triangulates against external reference series where they exist and are comparable.
- and labels as "directional" everything that is.
We would rather publish a figure with its limits stated than a figure with no base. When you cite these figures, state the base and the window that come with each one.
This methodology applies to the Panama panel. The Colombia panel will be documented separately, in line with DIAN electronic invoice structures and local data-protection requirements.
6How we differ from other sources
| Source | What it typically observes | Typical limitation |
|---|---|---|
| Shelf audit | displayed price, promotion and presence at audited locations | usually does not observe the final transaction, the full basket or the shopper |
| A retailer's sell-out | transactions recorded within that retailer | generally does not observe the same shopper's activity outside its stores |
| Web price tracking | publicly listed online prices | may not capture in-store prices, personalized discounts or the final paid amount |
| Cuestio panel | invoice-level transactions, paid prices and observed baskets across multiple submitted chains | opt-in coverage, no unqualified total-market estimates |
7Definitions
All figures on this page are cut as of August 3, 2026. They are computed on the consented, accepted base, unless the definition itself says otherwise. If we restate a published figure, we flag it in the delivery where it appeared.
| Term | Definition |
|---|---|
| Submitted record | What a user sends to the app: an electronic invoice scanned via QR or a manual expense entry. |
| Accepted record | A submitted record, either a DGI electronic invoice or a manual expense entry, that passed the validation and quality checks and entered the consented analytical base. At the cut: more than 400,000. |
| DGI electronic invoice | An invoice from the DGI electronic invoicing system, captured via QR and carrying a CUFE. At the cut: more than 85% of accepted records. All price, basket and line analytics run on these. |
| Manual entry | An expense typed in by the user. At the cut: less than 15%. It contributes no product lines and sits outside price and basket analytics by construction. |
| Purchase line | A product line extracted from an accepted electronic invoice, with description, quantity, paid price, discount and tax. At the cut: more than 1.6 million. |
| Product catalog | A normalized reference catalog of 54,000+ products with brand, category (25 categories) and barcode. Lines are linked to it where possible. |
| Mapped brand | A brand name normalized from purchase-line text and the catalog. One brand can group several written forms. |
| Merchant | An invoice issuer identified by its tax ID (RUC). At the cut: more than 9,000 distinct merchants observed in the panel's invoices. |
| Branch | A geolocated point of sale belonging to a merchant. One merchant can have several branches: the panel's merchants account for more than 14,000 geolocated branches. |
| Profile | A facturillo account with active consent. A profile is not a household: we never compare it directly with household panels. |
| Brand switching | A consented profile whose accepted purchases in a category move from one mapped brand to another between consecutive observation windows. Published only as aggregated counts over the section 4 floors. |
| Penetration | The share of active consented profiles in the window with at least one accepted purchase of the category or brand. |
| Reach | Distinct consented panel profiles with at least one qualifying purchase in the window. Always profiles, never households. |
| Velocity | Average qualifying transactions per active buying profile per time unit, stated with each report. |
8Frequently asked questions
Where does the data come from?
From fiscal documents that facturillo users submit voluntarily and that pass the acceptance checks. DGI electronic invoices can be checked by their CUFE or QR against the DGI system where lookup is available. Only profiles with active consent, and clients receive aggregated statistics only.
How large is the panel?
As of August 3, 2026: over 12,000 profiles with active consent. The base holds over 400,000 accepted records and over 1.6 million purchase lines extracted from the DGI electronic invoices. Every report states its definitions, its base and its time window.
Is it representative of all of Panama?
No. The panel is opt-in and is not a probability sample of the Panamanian population. It is built to measure prices, baskets, overlaps and trends observed within the panel, not to estimate unqualified market sizes. Every delivery states its base and its limits.
Do you sell personal data?
No. No client receives purchase histories or identifiers, direct or internal. Everything that leaves is aggregated, under the floor of 30 profiles and the suppression rules in section 4. Aggregates are built so they cannot be decomposed into the people who formed them. We publish curated editions, not open queries. Consent can be withdrawn anytime from the app.
How often does it update?
The facturillo app receives on average more than 45,000 submissions per month in 2026, scanned invoices and manual entries combined, and a share is accepted after the validation and quality checks. El Carrito Real is published monthly. Tailored reports state their data cut date.
Who is behind it?
Cuestio is the business-insights service built on consented, aggregated data from the panel of facturillo, Panama's finance app. Both are products of facturillo Inc., the US company behind the app. The facturillo app serves people and Cuestio serves companies, under one framework of consent, privacy and data governance.