Marketero

Marketero · Entrevistas

Ep 33: Clicks, Opens, Purchases, and Bot Traffic: What Email Platforms Aren’t Telling You | TWBERP

13 ideas de marketing sacadas de «Ep 33: Clicks, Opens, Purchases, and Bot Traffic: What Email Platforms Aren’t Telling You | TWBERP», de DTC Podcast…

Ver la entrevista entera en YouTube

De qué habla

Las 13 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 13 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.

Si confías solo en atribución de canal de Shopify, subinvertirás en email

Medición y experimentos

Shopify channel attribution está heavily biased hacia last-touch, que sesga hacia email como canal de cierre. Pero también es muy incompleto: no reporta todo. Los números de Shopify siempre van a estar por debajo de lo que email realmente produce. Usar solo esos números lleva a reducir presupuesto de email.

And I see email week in there. So, so I will just recommend that when we come to attributing, I would not use the Shopify channel attribution, get some software, get Google analytics, which is also got huge problems, get triple whale, or use your, use your platform revenues, take all your platform revenues and put them side by side to get an idea of what's happening. I mean, is it fine to have Shopify channel attribution as part of that mix? Sure. Sure. But do not rely on it. I guarantee you it is massively under reporting what's happening with your email and retention channel. And if you…

Escúchalo en el minuto 49:34

Net revenue de Shopify es el único número real: excluye cancelaciones y refunds

Medición y experimentos

Shopify muestra total revenue y net revenue. El net es el que realmente entró a tu cuenta después de cancelaciones y fulfiller issues. Cada plataforma reporta revenue inflado porque se atribuye 100% de una compra aunque múltiples canales hayan participado. Net revenue es tu ground truth.

That's like your total revenue. But of course, we all look at net revenue. So what's happening there? Well, look, did the order get fulfilled? Right? Because you see fulfilled order inside Shopify. And did the order get canceled? So when you pull out the stuff that didn't get fulfilled, stuff that, uh, stuff that got canceled, when you pull that out, you get your net revenue. So that's, let's just say that's the best revenue number. Someone could disagree, whatever. There's lots of different reasons to disagree with that. But let's just say the best revenue number is net revenue. Okay? That's…

Escúchalo en el minuto 44:29

Todo modelo de atribución es falso; algunos son útiles

Medición y experimentos

Klaviyo vs. Shopify vs. Google Analytics vs. Triple Whale dan números distintos porque cada uno tiene su propia lógica de atribución. La única métrica fiable es lo que realmente entró a tu cuenta: net revenue. Usa varios modelos en paralelo para tener una imagen aproximada.

Every attribution model is wrong and some attribution models are useful. You know, I mean only like Shopify net revenue is really the money you made. Everything else is just out to lunch. So sometimes I will want to just, Hey, say what we want is traffic. They don't know what's happening there. Yeah. Let's dig in and talk about what's a click, what's an open, what's a purchase. So you can understand the technology. You can understand your attribution. And you can troubleshoot problems when you see different systems giving you different numbers. You can understand what's happening behind that…

Escúchalo en el minuto 2:46

MTA attribution siempre subestima email: email tiene atribuida una fracción de la revenue que realmente genera

Medición y experimentos

MTA (multi-touch attribution) como Triple Whale divide la revenue entre todos los canales que participaron. En la primera compra con email + Facebook + Google, MTA le da 1/3 a cada uno. Pero si alguien compra 18 meses después solo por email, MTA le da 100% a email. El promedio sistemáticamente favorece a paid media.

So I have to attribute a hundred percent of that revenue to myself. And the only difference between platform attribution, and then we go to, let's just say triple whale, right? They will then have attribution logic that says, okay, but you know what? Sure. Email says, uh, it had, you know, that, that click and then sees an order, but it aggregates all that information. Yeah, but you know what? Gmail also had a click, right? And saw the open. And Facebook also had a click and saw the open, right? And it's not, it's not actually aggregating information from those platforms. It has their own…

Escúchalo en el minuto 18:09

Click, unique click y sesión son tres métricas distintas que se roll-up diferente

Medición y experimentos

Un click es una llamada HTTP. Una sesión agrupa clicks de un mismo usuario durante un periodo. Un unique click roll-up múltiples clicks de un mismo email en uno solo. Los clicks totais pueden duplicar las sesiones. No compares apples con oranges.

A click is a hyperlink that makes an HTTP call to a web server to serve HTML code. All right? That was technical, but I wanted to just give a decent definition. Now, also, that click could call an FTP server or an SMTP server. But for our purposes, when someone clicks, you know, HTTP colon forward slash WW whatever, right? So for most cases, it's just a hyperlink, which is a link that calls an HTTP server to serve HTML. And that HTML can have all kinds of stuff in it, like JavaScript, or it's been a long time since I've been a product manager. So I'm not even sure what's necessarily in sites…

Escúchalo en el minuto 3:19

El poder de email:免费 brand impressions durante 3 años que traen de vuelta sin costo de ads

Canales de adquisición

El valor real de email no es la primera compra sino mantener a alguien en la lista 2-3 años. Cada vez que llega un email de marca, refuerza el awareness aunque no hagan click. Cuando están listos para comprar, buscan la marca directamente. Ese tráfico no se rastreable en ads es lo que hace email免费的.

I mean, I've said this before, where the real money is made in email is in, in keeping someone involved with your brand for three years, right? And being there when they make their next purchase because stuff that's just in the funnel. I mean, sure. Like they get a checkout abandoned. Checkout abandons are huge. I'm not saying they're bad. They're amazing. The people end up back on the site and they purchase. Awesome. Right. But you also spent on Facebook and Google to get them on the site. And so, you know, only a, you can only attribute a portion of that in your mind to email. Email didn't…

Escúchalo en el minuto 18:09

Compara revenue de múltiples fuentes: Klaviyo platform revenue, Google Analytics, Triple Whale y Shopify net revenue

Medición y experimentos

Pon lado a lado los revenue reports de cada plataforma. No uses solo uno. Si la atribución de canal de Shopify es tu única fuente, subestimará sistemáticamente email porque usa last-touch. Corre las plataformas en paralelo y busca la tendencia, no el número absoluto.

And I see email week in there. So, so I will just recommend that when we come to attributing, I would not use the Shopify channel attribution, get some software, get Google analytics, which is also got huge problems, get triple whale, or use your, use your platform revenues, take all your platform revenues and put them side by side to get an idea of what's happening. I mean, is it fine to have Shopify channel attribution as part of that mix? Sure. Sure. But do not rely on it. I guarantee you it is massively under reporting what's happening with your email and retention channel. And if you…

Escúchalo en el minuto 49:17

Bot clicks inflan los click rates y sesgan los datos de email

Medición y experimentos

Sistemas de seguridad pre-cliquean links para verificar si el destino es seguro. Esto puede duplicar el click rate en ciertos emails. Los unique clicks mitigan esto porque roll-up múltiples clicks del mismo email en uno. Usa siempre unique click rate como benchmark.

So many systems out there will pre-click links to see what's happening on the other side. Right. And it's security stuff. Hey, hey, what is this? Is it safe for me to send this to my users? Right. When I click on this link, where does the, uh, HTML, where's the hyperlink call to? Like, oh, uh, a server in Russia. Right. You know, but it says it's in Wyoming. So, um, bot clicks have always been there. They were a big thing in 2024. Uh, and my guess is, you know, it's just becoming a bit less of a friendly world. Um, so, you know, there were just, I think probably a bunch of, um, corporations…

Escúchalo en el minuto 12:55

Delivery rate no mide inboxing: el 99% de entregabilidad solo significa que no rebotó

Medición y experimentos

Cuando ves 99% de delivery, eso solo dice que el email llegó al servidor del proveedor. Puede haber ido a spam. Usa GlockApps o herramientas similares para medir inboxing real. Delivery rate es la métrica más engañosa del email marketing.

And then you, we didn't discuss it, but then you have bounces, bounces where it is where it just doesn't get there. I guess let me just take a second. A hard bounce address doesn't exist. So some server says that address doesn't exist. These are the worst. And there's soft bounces, which is often like the mailbox is full or all kinds of things can cause soft bounces. But the point is you could possibly send to that email address in future and have it received. Whereas with a hard bounce, it'll never be received. The address doesn't exist. After that, that's message received, right? Now here's…

Escúchalo en el minuto 38:52

La información de email marketing es inherentemente con pérdida (lossy)

Medición y experimentos

HTTP calls fallan, Shopify no registra todo, los pixels de apertura se cachean con Apple Privacy. Cada paso de tracking puede fallar independientemente. Por eso Klaviyo muestra visits sin clicks: alguien buscó tu dominio directo en el navegador sin haber abierto el email.

So these HTTP calls, first of all, they can fail. You can obviously have, they, can fail in that they, fail to load anything. Cause you're calling something that's not there. Like it's a bad link. They can fail because for some reason the server fails. So even though there was something there, you know, nothing happens. It just spins for a while. All kinds of things can go wrong. So understand that you can have clicks where it works, but doesn't get tracked for some reason. Cause, the tracking thing didn't work. So this information is inherently lossy, right? And so lossy that I, you know, I…

Escúchalo en el minuto 10:05

Apple Privacy hace que los open rates estén completamente distorsionados

Medición y experimentos

El 90% de usuarios de iPhone dice sí a Apple Privacy. El pixel de tracking se cachea en un servidor de Apple, y se carga automáticamente aunque no abras el email. Esto infla los open rates. Un open rate del 50% puede ser completamente falso.

someone opens up their, a new iPhone, it asks, do you want to protect your identity? 90% of people say yes. 10% of people, for some reason, don't want to protect their identity. And those people, when, when an email is opened in an email application, the tracking pixel that we discussed is cached on an Apple server. They do that to protect your, for a number of reasons, one, to protect your location. So you can't be geo-located with the loading of that, with the, with the pixel, with the call. Um, because all along the line, as during that call happens, there are, you go through routers and…

Escúchalo en el minuto 28:49

La atribución por defecto de Klaviyo (5-day open / 1-day click) refleja mejor el rol de marca de email que 7-day click

Retención y relación con el cliente

7-day click optimiza hacia CRO: solo creditea cuando hubo click. Default attribution también creditea opens, lo que captura el impacto de marca de mantener a los clientes en contacto durante años. Si cambias a 7-day click, tu revenue reportado baja pero pierdes visibilidad del efecto retención de marca.

order, but Klaviyo default attribution is if they've opened in five days or clicked in one day. And what people often do on other platforms is do like a seven day click. What's right. What's wrong. Every attribution model is wrong and some are useful. So a seven day click of course means one of these clicks was registered by the Klaviyo a code that's in the site and, and then Klaviyo sees from that user. Cause that click is recorded back to the user. Klaviyo sees a purchase that happens within seven days for that user records that revenue. It attributes that revenue. The default attribution…

Escúchalo en el minuto 15:47

Usa read rate de Litmus (8+ segundos) como proxy del engagement real en email

Medición y experimentos

Read rate mide cuánto tiempo alguien pasa en el email. Un read en Litmus son 8 segundos. Esto filtra los Apple Privacy opens falsos y te da una idea más real de quién realmente leyó el mensaje. Puedes usar la ratio de reads/apple-privacy-opens para estimar engagement real en toda la lista.

So then there's this thing called read rates and probably very few of you use these, but they are pretty cool. I've used read rates from litmus. And what it does, I'm actually not entirely sure what's happening on the technology. I think, I think they probably have, they leave sessions open. So a session can have an expiry date, you know, the session in your browser, expiry time. And it's often like instantaneous, but I think they leave sessions open. And so they're able to track how long that, when that session gets closed. Bottom line, they can track how long someone was in your email…

Escúchalo en el minuto 26:29

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.