Ellos dicen
«Invoca already attributes calls to sources.»
Nosotros decimos
It does, well. What it cannot see is the invoice, so it cannot say which of those sources produced revenue.
Para Invoca
A call classified as qualified is a judgement about a conversation.
Export your call records, match them against closed sales, and find out which sources produced calls that became invoices.
Sin clave de API Nada que instalar en InvocaNada que instalar Sin tarjeta de créditoSin tarjeta
Invoca sits on the lead side in CloseRev: export its call records — the caller's number, the source that produced the call and the date — and match them against closed sales from whichever system holds them. Invoca's own strength is classifying what happened during a call; what it cannot see is whether the business was ever invoiced, which is the question a marketing budget turns on. Revenue with no traceable lead is reported as Direct / Unknown.
Comprobado por última vez con la documentación del propio Invoca el 24 de septiembre de 2026.
La brecha
You have excellent classification of what happened on the phone, and no idea which of those calls the company was eventually paid for.
Invoca ve
Qué se vendió, a quién y por cuánto.
ninguna fila en común
Su cuenta publicitaria ve
El clic, la palabra clave, la llamada y cuánto costó cada uno.
El archivo
Qué debe contener la exportación.
Tres cosas sostienen la coincidencia: quién, cuánto y cuándo. Todo lo demás es opcional y solo cambia cómo se puede desglosar el informe.
- Necesaria
Caller's phone number
On the call record. This is the strongest lead key in the whole catalogue, because it is the number the person actually used.
- Necesaria
Source, campaign and keyword
Whatever the record carries about what produced the call. This is the channel label the revenue will be grouped by.
- Necesaria
Call date
When the call happened, which anchors the window.
- Opcional
Classification or outcome
Qualified, booked, not a fit. Keeping it lets the report compare the classification against what the ledger later said.
Paso a paso
Cómo sacar el archivo de Invoca.
Escrito para quien tiene Invoca abierto en la pestaña de al lado. Los nombres de los informes varían según la edición, así que cada paso indica qué buscar.
-
Export the call records
One row per call with the caller's number, the source and the date.
-
Keep the classification
It is the thing worth testing. Comparing it with closed revenue tells you how well the model predicts money, which nobody can know from inside it.
-
Export closed sales
A contact detail, the amount and the date. Collected revenue rather than booked, where the two differ.
-
Include the calls that were not qualified
Some of them closed. A file containing only the qualified ones cannot discover that, and that discovery is usually worth the exercise on its own.
-
Upload both
The join runs on the phone number, normalised, with unmatched calls and unmatched revenue both kept visible.
Lo que recibe
La página que Invoca no puede mostrarle.
Ingresos por canal, el número de ventas detrás de cada cifra y un grupo honesto para las que nadie pudo rastrear. Cifras de ejemplo, del caso práctico de la página Redes hospitalarias, no de una cuenta de Invoca.
Atribuidos a un canal$1,715,00049% de $3,500,000
Ventas coincidentes411 de 813solo alta confianza
Ticket medio$4,400por venta pagada
| Canal | Proporción | Ventas | Ingresos |
|---|---|---|---|
| Google Ads | 214 | $1,015,000 | |
| Meta Ads | 118 | $420,000 | |
| Email marketing | 79 | $280,000 | |
| Directo / Desconocido | 402 | $1,785,000 |
Las ventas sin coincidencia quedan en Directo / Desconocido. Nunca se reparten entre los canales pagados para que el total se vea mejor.
El argumento
Qué cambia cuando cada call y el anuncio comparten una fila.
A classification is a prediction and the ledger is the outcome
Scoring a call as qualified is a judgement made from the conversation, and it is usually a good one. It is still a judgement made before anybody paid anything.
Comparing classifications with closed revenue turns that judgement into something testable: what share of qualified calls produced an invoice, and what share of unqualified ones did too.
Both halves are informative. A high rate in the first is confirmation; anything meaningful in the second is a rule quietly discarding business.
This is not an argument against classification. It is the check that tells you whether it is calibrated for your business, which cannot be established from inside the tool that produces it.
The phone number is the best key anybody in this catalogue has
Most lead files hold a number somebody typed into a form, which can be mistyped, deliberately wrong, or a work number for a personal purchase.
A call record holds the number the person actually called from, verified by the act of calling. It matches against a customer record more reliably than anything else available.
That makes this one of the highest-yield sides of a match, and it is worth exporting even when another system is your primary lead source.
The usual failure is on the other side: the revenue system holds an old landline or a relative's number. Normalisation runs on both, and what does not join is reported rather than dropped.
Source detail here is finer than most channels can supply
Call records of this kind routinely carry the campaign and often the search term that produced the call, which is a level of detail most lead files never reach.
Grouping closed revenue at that level shows which terms produce calls that become invoices rather than calls that merely happen, and the two lists are rarely the same.
It is also the level at which a large account can act without restructuring anything: pause terms, shift budget, change the offer behind the ones that work.
Where the volume at term level is too thin to be stable, the report rolls up rather than printing a ratio built on four calls.
Calls nobody answered are a finding, not a gap in the data
Every call file contains rings that went nowhere — after hours, queued too long, abandoned — and they are usually excluded from marketing reports as noise.
They are the clearest cost in the file. A source producing excellent calls at a time nobody is on the phone is being judged for an operational failure.
Keeping them in and reporting them separately puts a revenue number on the problem, which is the form in which it tends to get fixed.
It also protects a channel from being cut for the wrong reason, which is the most expensive mistake available in this whole exercise.
Preguntas justas
«Invoca ya hace eso». No exactamente.
Ellos dicen
«Our classification is accurate.»
Nosotros decimos
Probably. This is how you demonstrate it rather than assert it, and how you find the cases it is quietly missing.
Ellos dicen
«We only care about qualified calls.»
Nosotros decimos
Then export the others anyway once. Whatever share of them closed is business a rule is currently discarding.
No. It works from an exported call file, so no recordings, transcripts or live call data are reachable from here.
Call records with the caller's number, the source and the date; closed sales with a contact detail, an amount and a date.
No, and they should not be exported. The number, the source and the date are the whole requirement.
Yes, at least once. Some of them closed, and finding that out is often the most valuable result.
Yes — that is the point of keeping the outcome column. It compares a prediction against what was invoiced.
As fine as the record carries, often to the search term. The report rolls up where the volume is too thin to be stable.
Reported separately, with the revenue that never followed them, because that is usually an operations problem rather than a channel one.
The phone number, normalised — country code, punctuation, extensions — because it is the most reliable key available.
Both sides are normalised and the email is used as a second key where present. Anything that does not join is reported.
No. It adds the half that happens after the call ends.
A phone number, a source and a date on one side; a contact detail, an amount and a date on the other. No recordings, no transcripts, no conversation data. Encrypted in transit and at rest and deleted with the import.
Closed revenue by source, campaign and term where available, classification tested against outcome, missed calls reported apart, and everything unmatched kept visible.
Por sector
Sectores que usan Invoca.
Cómo se ve el informe una vez cargada la exportación, escrito para cada uno.
- Redes hospitalarias¿Qué canales llenaron las líneas de servicio?
- Seguros comerciales¿Qué canales generaron pólizas emitidas?
- Cooperativas de crédito¿Qué canales generaron cuentas y préstamos financiados?
- Servicios para el hogar¿Qué anuncios produjeron trabajos, y no solo llamadas?
- Residencias para mayores¿Qué canales produjeron admisiones y cuánto generaron?
Otros sistemas
También vale la pena leer.
¿Usa más de un sistema o está comparando? El método es el mismo; las columnas, no.
Archivo de leads y llamadas CallRail Empresas que cierran sus ventas por teléfono Archivo de leads y llamadas Qué exportar Archivo de leads y llamadas CallTrackingMetrics Teams whose leads arrive by phone and by form Archivo de leads y llamadas Qué exportar Archivo de leads y llamadas Google Ads Search advertisers Archivo de leads y llamadas Qué exportar
Los 97 sistemas 
Invoca y los demás nombres de productos y logotipos de esta página pertenecen a sus propietarios y se muestran para identificar el software del que procede un archivo. CloseRev no está afiliada a ellos ni cuenta con su respaldo, y no se conecta a ninguno: lee un archivo que usted exporta.
Empiece hoy
Dos exportaciones y lo sabrá.
Nada que instalar en Invoca, sin clave de API y sin necesidad de haber rastreado nada hasta ahora. El año pasado funciona igual de bien que este mes.