Marketero · Entrevistas
How Meta Ads Are Changing Forever: Advantage+ & cAPI Subterfuge REVEALED!
25 ideas de marketing sacadas de «How Meta Ads Are Changing Forever: Advantage+ & cAPI Subterfuge REVEALED!», de Perpetual Traffic. Sobre todo de canales…
- 25 ideas despiezadas
- Fuente Perpetual Traffic
- Dura 1 h 00 min
- Publicada en 2025
Ver la entrevista entera en YouTube
De qué habla
Las 25 ideas de esta entrevista se reparten así entre las etapas del marketing. El reparto es lo que distingue una entrevista de otra: dice en qué se detuvo de verdad quien hablaba.
Las 25 ideas, con su cita y su minuto
Cada una está sacada de un punto concreto de la entrevista. El enlace del titular lleva a
la página de esa idea, donde se ve quién más la dice por separado; el del pie salta al
segundo exacto del vídeo para que la compruebes tú.
Estructura publicitaria consolidada: una campaña, un ad set y múltiples creatividades
Canales de adquisiciónPara operar con los nuevos algoritmos de Meta, la estructura recomendada consiste en consolidar en una única campaña con un solo conjunto de anuncios (ad set) que contenga múltiples variaciones creativas diseñadas para la parte alta, media y baja del embudo.
And so what meta has kind of released to us in Iceland last week was the way that they are ultimately moving to is one campaign, one ad set when possible, and then multiple pieces of creative from top, middle and bottom of the funnel,
Escúchalo en el minuto 19:48
Las inclusiones y exclusiones de audiencias personalizadas en Meta Advantage+ operan como sugerencias probabilísticas y no como reglas estrictas
Canales de adquisiciónEn las campañas con Advantage+ Audiences de Meta, configurar exclusiones de clientes existentes o inclusiones de audiencias personalizadas no garantiza un filtrado estricto. El algoritmo trata estas audiencias como sugerencias ('probabilísticas, no deterministas') y prioriza la optimización y segmentación a través de los propios creativos del anuncio.
And then a custom audience exclusion that is now basically used also with the vantage plus audiences as a suggestion. This is now considered a suggestion. I can still exclude my existing customers. No, it is a suggestion. I dictate by ad by the creative now. what's interesting here is you see how I can suggest an audience and that's where it looks like custom audience inclusions look very similar to custom audience exclusions. They're still all suggestions. As meta said, it is probabilistic, not deterministic was their verbiage. So your inclusions and exclusions are probabilistic, not…
Escúchalo en el minuto 40:19
El algoritmo de deduplicación de Meta y la segmentación guiada por creatividades
Canales de adquisiciónMeta implementó un algoritmo de deduplicación de oportunidades entre campañas y conjuntos de anuncios que comparten eventos de conversión. Como consecuencia, las audiencias tradicionales pierden relevancia y la segmentación efectiva queda determinada por el copy, las creatividades y la página de destino.
And the team that I've been working with inside of meta has explained the opportunity de-dupe ad set algorithm that just went live. The de-duping algorithm that just went live essentially means that when an opportunity is found in a same conversion event as in two different campaigns and or ad sets, meta will now de-dupe that opportunity into one. And then you do not get to choose or identify which one that is. And now it is prospecting based on your words, phrases, your creative, the interpretation of that creative and the landing page that you're sending those users to. Now that is now your…
Escúchalo en el minuto 10:02
La transición obligatoria a Advantage+ en Meta delega la segmentación al contenido creativo
Canales de adquisiciónMeta Ads está eliminando progresivamente la opción de configuración manual de campañas y la segmentación granular de audiencias, forzando el uso de Advantage+ donde la identificación del público objetivo depende del algoritmo guiado por el propio anuncio y su creatividad.
And now this is where you could do a manual setup for your manual sales campaign. you're probably not going to see that manual setup part. Watch what happens when I hit create sales continue. It bypasses it. It is no longer there. And that's what we saw is the future. this is introduced gradually. You're on in your advantage plus campaign. where's my targeting, don't worry. Upload a video of a dog and they'll find. Don't worry. We'll figure it out for you.
Escúchalo en el minuto 36:18
Optimizar campañas de Meta enviando solo conversiones verificadas y segmentadas vía CAPI
Medición y experimentosPara evitar que los algoritmos de Meta se atribuyan ventas incorrectas u optimicen con datos distorsionados, se debe alimentar la API de conversiones (CAPI) exclusivamente con eventos personalizados basados en datos de primer clic. Esto implica acreditar a la campaña únicamente si la conversión proviene de un cliente nuevo o recurrente (según el objetivo de la campaña) y si se vendió exactamente el producto…
That means we get to import the conversions from our own data. We essentially remove the pixels capability of feeding itself. And now we feed the pixel the data we want the pixel to optimize for. the click is actually coming from your campaign. you do not get to take credit for it. if we're trying to sell a water bottle and you sell a shoe, you now do not get credit for it. If you sell a returning customer of one of ours, you no longer get to take credit for it. And it has to be the product that's actually sold in that campaign, If you haven't verified that the product you're trying to sell…
Escúchalo en el minuto 30:35
La creatividad y el contenido actúan como el mecanismo principal de segmentación en Meta
Canales de adquisiciónMeta ha eliminado progresivamente la segmentación manual por intereses, demografía y psicografía, dejando únicamente la geográfica activa y convirtiendo a la creatividad y al contenido del anuncio en el factor determinante que guía a la IA para seleccionar la audiencia.
And their targeting is now based on content is no longer based on audiences and interests and segmentations. And psychographics and demographics still a geographics, but that's the only one that's left. So your content, or your creative really, The creative.
Escúchalo en el minuto 19:29
Optimizar campañas explícitamente por nuevos clientes para evitar canibalización por recompras
Canales de adquisiciónConfigurar las plataformas publicitarias (Meta y Google Ads) para optimizar específicamente por adquisición de nuevos clientes (NCA) en lugar de compras generales previene que el algoritmo asigne presupuesto a usuarios existentes que ya iban a recomprar.
but we've been still seeing this when we increase our costs on everything outside of a NC, a new customer version, we ended up getting a lot of existing customer growth. So much so that we even saw this inside of Google ads where the new customer imports are finally actually starting to grow in a good sense where we're actually seeing more new customers and returning customers. but the way that we're fighting it against is going after those per purchase and sees, but this can become very, But what was very interesting is whenever we didn't ask for a new subscriber, we saw that it was like 53…
Escúchalo en el minuto 22:58
Advantage+ de Meta prioriza volumen de conversión sobre segmentación y puede canibalizar otros canales
Canales de adquisiciónLas campañas Advantage+ de Meta ignoran las restricciones tradicionales de segmentación y exclusión para perseguir el evento de conversión asignado (como la compra estándar). Debido a esto, el algoritmo tiende a atribuirse conversiones de usuarios preexistentes o generados por otros canales (email marketing, SEO, Google Ads o tráfico de influencers), mostrando métricas infladas dentro de la plataforma.
Advantage plus now says, I'm going to go after anybody I want, regardless of what you say, regardless of your targeting, regardless of your exclusions, I'm going to go do whatever I want, So we have the advantage plus that basically says, I met a do not care advertiser. What you want to go after. All you can tell me what to do is get a conversion. So meta says, because I can go out to purchase. You are allowing me to go after traffic that I didn't generate from emails that maybe from a year ago, from organic traffic that you're already paying your SEO team for, and Google traffic that you're…
Escúchalo en el minuto 26:09
Medición y experimentosSi no se puede registrar el ciclo de conversión con alta precisión (por ejemplo, perdiendo los primeros días de un ciclo de ventas largo), importar datos propios o eventos vía CAPI a Meta o Google resulta contraproducente. Es preferible confiar en la optimización nativa de las plataformas antes que alimentar sus algoritmos con datos parciales o sesgados.
So when you're looking at where your actual conversions, the source of those conversions are, if you're not tracking this with a high, high degree of accuracy, if you're missing the first five days of a 30 day sales cycle, your imported data is junk. If you're missing the first five days of a 30 day sales cycle, that data is junk and you're only going to hurt yourself. It's better off just letting Google or meta do it.
Escúchalo en el minuto 35:08
Captar clientes B2B en Google Ads mediante búsqueda con palabras clave de intención mayorista
Canales de adquisiciónPara aislar la demanda B2B del tráfico retail B2C en Google Ads, se debe priorizar campañas de búsqueda (Search) en lugar de Shopping y pujar por términos que indiquen intención corporativa explícita (como «distributor», «manufacturer» o «mass bulk»).
the best thing that I've seen is an inbound search using different variations of keywords that would show intent for a company that they're trying to look for things like best cleaning manufacturer distributor, the keywords that are, they're looking not to buy at retail, but they're looking to come to the company to do so. but if you're looking for B2B, do search, not shopping and show up for keywords that are things like mass bulk, manufacturer distributor, those types of things will then bypass the thing like, that's a whole different intent. And that could help you get around the B2C…
Escúchalo en el minuto 55:54
Medición y experimentosDe forma análoga a los tests A/B en email marketing, conviene probar nuevos anuncios asignando una fracción del presupuesto publicitario a una muestra reducida de la audiencia y, una vez confirmados los creativos ganadores, incorporarlos a la campaña principal de gran escala.
but this goes back to like email marketing testing, which means I'm going to take one third of my ad spend, do creative testing on this when it's good. It stays there. But now these are added to my larger campaign that spends more. So if it works good with a small one, it's almost like testing your ad headline before a massive blast. Give me 1,000. I'm going to do some creative testing. While it works, release to the masses.
Escúchalo en el minuto 54:11
Medición y experimentosDe forma análoga a los tests A/B en email marketing, conviene probar nuevos anuncios asignando una fracción del presupuesto publicitario a una muestra reducida de la audiencia y, una vez confirmados los creativos ganadores, incorporarlos a la campaña principal de gran escala.
but this goes back to like email marketing testing, which means I'm going to take one third of my ad spend, do creative testing on this when it's good. It stays there. But now these are added to my larger campaign that spends more. So if it works good with a small one, it's almost like testing your ad headline before a massive blast. Give me 1,000. I'm going to do some creative testing. While it works, release to the masses.
Escúchalo en el minuto 54:11
La automatización de Meta favorece estructuras simplificadas y penaliza la compra técnica manual
Canales de adquisiciónMeta ha consolidado sus formatos automáticos (como el paso de Advantage+ Shopping a Advantage+ Sales como opción por defecto), restando efectividad a las exclusiones manuales y a modelos como ABO (Ad Set Budget Optimization). El algoritmo optimiza las conversiones de forma eficaz en estructuras simplificadas, pero a costa de reducir el control del anunciante sobre el origen y tipo de cliente adquirido.
We now have a new update to advantage plus shopping is now called advantage And this now defaults into each meta advertisers account. Essentially what this means is that what we were kind of whistleblowing for like the last three months is that it does not look like any of our targeting or exclusions are working. My AI is smarter than your technical media buying. The problem is it is so powerful that the technical media buying that has now gone away, that has been replaced by AI is in fact so powerful that it can attribute and contribute conversions extremely well. The problem is it may be on…
Escúchalo en el minuto 6:32
Medición y experimentosEnviar conversiones directas de Shopify a través de Conversions API (CAPI) basadas en último clic a campañas de prospección descalibra la optimización algorítmica, ya que la atribución predeterminada de Shopify maneja solo un 44% de precisión sobre el origen real de la compra.
Being able to first click. Cause you Cappy import last click on a first click campaign. So first click Cappy import of true first data is now King. So if you're doing like your Shopify imports, like if you have a direct integration with Shopify and you're sending in your Shopify purchases, even though Shopify is actually measuring 44% accuracy based on even Shopify's public data. They say that our attribution miles is only 44% accurate. You're going to Cappy import that. And you're going to be 56% wrong.
Escúchalo en el minuto 22:22
Medición y experimentosNo se debe asignar el mismo objetivo de costo de adquisición de clientes nuevos (nCAC) o límite de costo (cost cap) a productos con distintos márgenes brutos; aquellos con mayor rentabilidad permiten pujar más alto para captar clientes de manera rentable.
But if you're selling it individually, product A at a 70% gross profit, you should not have the same NCAC as product B with a 40% gross profit. It just doesn't make sense. why would you cost cap something that you have greater profitability on when you know that you can afford to acquire that customer at a greater price?
Escúchalo en el minuto 50:21
Medición y experimentosEs necesario calcular y optimizar el costo de adquisición de nuevos clientes (nCAC) a nivel de cada producto individual en lugar de manejar un nCAC global promedio, ya que los márgenes de rentabilidad bruta pueden variar drásticamente entre productos de un mismo catálogo (por ejemplo, del 40% al 70%).
Is that this now goes from a global NCAC to a product-based NCAC. Because your NCAC per product varies greatly. So it's a basket of products, about six different products. Their gross profitability varies from 70% down to 40% based upon the individual product. The global NCAC for that basket is X.
Escúchalo en el minuto 49:46
El límite de escalar dependiendo exclusivamente de creatividades transaccionales de fondo de embudo
Mensaje y copywritingDepender únicamente de anuncios transaccionales de fondo de embudo termina limitando la escala de adquisición; para crecer es necesario diversificar creatividades incorporando ganchos diferenciadores, historia de marca/fundador y propuestas de valor únicas.
what's so funny is we're not even drinking our own Kool-Aid. We've been doing all bottom of the funnel and that's where I think we're limited. This last week, we are changing our creative strategy to produce more creative. Cause we're like 10 bottom of the funnel ads. Now we're like, We have founder story, unique hook, like why buy your turkey tail mushroom from us versus anybody else? What makes you special? That needs to be in there. So we're actually expanding our creative a lot in order to test this.
Escúchalo en el minuto 57:46
Escala y referencias de puntuación EMQ en la API de Conversiones de Meta
Medición y experimentosLa métrica de Event Match Quality (EMQ) de Meta tiene una puntuación máxima perfecta de 10, considerándose valores de 8 y 9 como resultados de alta calidad para la coincidencia de datos de eventos enviados.
And those are EMQ scores. A perfect match rate for EMQ is 10. Obviously nines and eights are fabulous.
Escúchalo en el minuto 29:42
Segmentar datos vía CAPI para dirigir la optimización hacia clientes nuevos
Medición y experimentosDado que las campañas automatizadas de Meta tienden a tomar el camino más fácil y convertir a clientes existentes si no se restringen, una alternativa técnica para recuperar el control de adquisición es importar eventos mediante la API de Conversiones (CAPI) diferenciando explícitamente entre clientes nuevos y recurrentes a nivel de producto.
Our solve for that was to do cappy imports by new customers, returning customers, the product, new customer by product, the returning customer. I guess I would say into regaining some sort of control so that we don't end up in a situation where you launch just a simple advantage plus sales campaign. that has been replaced by AI is in fact so powerful that it can attribute and contribute conversions extremely well. The problem is it may be on conversions you don't necessarily want. So we've been running some import tests and those import tests that we have been actually now utilizing the new…
Escúchalo en el minuto 6:53
Impacto en NCAC al controlar la adquisición de nuevos clientes en Meta
Medición y experimentosAl enviar datos precisos de conversión para restringir la optimización a primeros compradores, la proporción de nuevos clientes subió del ~46% (16 de 35 ventas) al 84% (42 de 50 ventas), reduciendo el coste de adquisición de nuevos clientes (NCAC) de 219 USD a 203 USD.
And we have 35 sales. 32 of them are the product that's being sold. However, only 16 were new. 50 sales, 42 new customers, 42 new customers of your immune. Perfect. We've increased this from a 50% to now like an 85% where 42 out of 50. Now we're getting somewhere a 80% increase in the amount of new customers We spent 35, 17, 35, 17. Divide that by the only 16 amount of new customers that came in. We're spending 219. This one is already beating it at 203. This one got $219 NCAC. This was getting $203 NCAC.
Escúchalo en el minuto 13:16
Las mejoras de CPA (-24%) y ROAS (+22%) en Advantage+ pueden inflarse por atribución de marca
Medición y experimentosAunque Meta reporta una reducción del 24% en CPA y un aumento del 22% en ROAS con Advantage+, estos resultados pueden estar sesgados debido a que el algoritmo prioriza conversiones fáciles de bajo costo como búsquedas de marca y retargeting de último clic en lugar de prospección incremental real.
But the thing that we notice most often is that the metric that meta quoted on advantage plus compared to manual, 85% of advertisers out there that had this turned on without them understanding or knowing or agreeing to it, which means their campaign started using advantage plus audiences in manual campaigns. There are metrics that they went to and said, this is the results that we'd like is a 24% reduction in CPA and a 22% increase in ROAS. If you target people that may not want you to target and you claim credit for them anyway, your metrics look good. My business hurts. It goes after…
Escúchalo en el minuto 17:52
La optimización de nuevos clientes en Google Ads es probabilística y no una exclusión estricta
Canales de adquisiciónEl objetivo de adquisición de nuevos clientes (NCA) de Google Ads opera como una optimización probabilística fundamentada en historiales y actividad de compra, no como una regla determinista o una exclusión estricta de clientes existentes.
Google's new customer acquisition optimization versus strict exclusions. Google's NCA goal, even when used is an optimization strategy, not a strict only show as new customer switch. So now the algorithm aims to find users. It believes our new customers bases based on Google's data, things like search history and purchase activity. probabilistic, not deterministic.
Escúchalo en el minuto 45:21
Canales de adquisiciónAl lanzar una nueva marca o campaña de un producto con alta demanda comercial preexistente (como ropa o calzado), conviene comenzar captando a los usuarios con alta intención de compra inmediata (bottom-of-funnel) antes de invertir en generar concienciación en la parte alta del embudo.
I always like the bottom up funnel, like three years ago is a bottom up funnel strategy where I said, start with people who are going to buy. If you have something that people are going to buy, like an established in stream shoes, t-shirt hats, whatever it is, those people that want that now start there and then work. Work your way up. So don't try to figure out how to get people interested when they are ready to buy from you. There are already people ready to buy from you. Work your way up. Catch the low hanging fruit first before you go to the top of the
Escúchalo en el minuto 56:54
Dinámica de embudo dentro del ad set en Meta y su impacto en el testeo creativo
Canales de adquisiciónAl agrupar creatividades secuenciales dentro de un mismo conjunto de anuncios en Meta, el primer anuncio (como la historia del fundador) suele absorber la mayor parte del gasto y clics con un rendimiento directo bajo, pero empuja a los usuarios hacia los anuncios de medio y fondo de embudo que consiguen CPA bajo y alto ROAS.
But the way that meta set is be mindful of when you do that, your creative testing is no longer creative testing in the traditional sense. What they say is that you're going to find something like the founder story that is actually the first ad in the ad set is going to have a large amount of spend and clicks, but very poor performance. But the longer it goes on, your middle and your bottom of the funnel will have less clicks and really low CPA and really high ROAS because it is pushing it through a funnel, which is why they're locking an audience into the ad set.
Escúchalo en el minuto 20:29
Reducción del 9% en CAC de primeros compradores al escalar gasto de $25k a $1M+
Medición y experimentosAl escalar la inversión publicitaria de $25,000 a más de $1,000,000 de dólares, el costo para adquirir clientes primerizos disminuyó un 9%.
Like that's what we put in. We went from $25,000 to over a million dollars in spent. Our cost of acquired first time customer went down 9%.
Escúchalo en el minuto 51:15
Seguir leyendo
Esta es una de las 1872 entrevistas despiezadas. Puedes
ver todas, leer las ideas por tema o
empezar por la ruta guiada si no sabes por dónde.