Review 1 - Digital analytics Minidegree

Hello all, I am thrilled to say that I have got into digital analytics minidegree with the amazing cxl institute, and the reason of that is any digital marketer needs to be brilliant and smart when it comes to data, because all of your decisions based on it

When i enrolled and got accepted i checked all the courses’ info, and unfortunately I could not resist my desire to start with attribution course by Russel Mcathy

For me it is the most interesting part in analytics world, because it is getting bigger and bigger and we execute a lot every single day and we need to know what step deserve what credit

Mcathy started his intro by showing the 4 characteristics for attribution models theoretical, technical, statically and tactical but first what is attribution and why do we need it ?

In football we do not give all the credit to the one who scores the goal, and the same in attribution, attribution is all about modules shows how to distribute the conversion credit through all the buyer journey, because it is about understanding the consumer, So we need to understand the full journey and the full value of the way that the consumer engages with the brand, and this consumer journey has to have an end point, And so what we're going to be calling a conversion today is that end point.

You must  understand the context behind the data that we're able to get, We have to put performance in the context of everything that goes on around what marketing is.

Attributing throughout a what is the true performance of all of our activity through to an ultimate conversion, whether that conversion happened in store, through the POS data, or actually through the digital journeys online. Whether that conversion happened in store, through the POS data, or actually through the digital journeys online.

How do we actually engage with people? And with what frequency ?

There are two types of companies, high frequent low value and high value low frequent, the first one like clothes store, and second one like cars store or air fly tickets

With high frequent project it takes 30 days for engage and re- engage, and with low frequent it takes up to 5 years for rethinking to change or re-use

And based on some recent researches the conversion happens after up to 4 visits in ecommerce

What are attribution models ?

• Last click : gives the credit to last click event
• Non direct last click : gives conversion credit for pre last click event
• Linear : equally credit, it doesn't show marketer what channel impacts the user
• Time decay : when the conversion happens, let's redistribute the value of that conversion back through previous historic visits, But we're going to scale that back over time, take into consideration the amount of time between each time
• Position based : the first click and the last click have the most credit

What data do you need to know for an attribution solution ?

are you able to get event level data to the level that you need?, Are you able to see not just a visit?, Are you able to see things within that visit?, Are you able to see the page views?, What the URLs are?, That'll be really important to understand the journey and the interactions within it. Are you able to look at things that happen within that page whether that be a JavaScript event or a change in the DOM?, Are you going to to be able to track that and say, when that happens, that gives us an indication that that consumer has changed their decision making process?, Are you able to see everything that goes on within that website, whether that be in an Ajax form or whether that be a drop down menu and the interactions within it?, Are those things important to you?, Are you able to see the device the person is using, their location, etcetera?

There are two different type of events : 

• Macro event
• Ecommerce sale, lead generation
• Micro event 
• Product view, add to cart, blog post view

there is going to be a significant amount of micros within that journey. A micro can be a categorization of a multitude of different pages. So that micro could be that, that user has viewed category page. At that point we don't need to know what category page it was but the fact that that user has got to that level within the process is the thing that we're trying to trigger and that's what we want to show to the attribution model. That user then does a site search. We don't necessarily need to know what they searched for at this moment but it would be good to have that understanding.

Attribution is going to take you on to the next level and that's the focus here, we know with the data is not going to be 100% correct. What's important for you is understanding that you get the best picture you can and the most amount of information you cannot make smarter decisions, because that's when you end up as a smarter marketer.

Tactics 

1 - CRO

Conversion rate optimization typically focuses on improving the conversion rate on a singular interaction. Whether that be a landing page or a page within the journey or a booking flow

when it comes to conversion optimization and you optimize a landing page we're not just optimizing towards that landing page taking someone to the following page and even in an advanced CRO output, you're also looking at the advancement from that landing page through to the final conversion.

The data that is made available through attribution is going to allow you to say this was the landing page and this is how it impacted all future interactions moving forward

Segment that by different user types and then use that to understand what value the conversion rate optimization activities you're doing are impacting those groups of individuals over a longer period of time.

2 - PPC

a lot of the activity you're going to be doing is going to be very different, whether it's brand or generic.

Brand for example is typically right at the end of a journey. It will have a low cost per click it will have a high CTR, Those types of users are going to have very high brand awareness and exactly the type of user who you want to convert very quickly. What that actually means is that they will come to the website and within one or two visits will be converting.

If that person is engaging with you on a generic PPC term, he tend to not have as much knowledge of the brand or service that you're offering.

What you can do there is you can look at the attribution value of him separately and understand actually what the consumer is waiting at that point in the journey.

You don't want user to buy today, What you want to do is increase their likelihood to buy in the future

The first visit could have been generic PPC where someone didn't really know what they wanted to search for. 

They were looking for research at that time.

The last visit may be brand PPC where the person was ready to buy from that particular company.

By default, in most attribution models where you're focusing on the last interactions you're going to be giving the majority of the value if not all of the value to the last interaction and generic PPC will get no value at all.

This is why we tend to see a much higher CPC and CPA on generic campaigns.

These are people who are not going to convert in that visit the majority of the time

You're not able to use the same logic and same management for a generic campaign as you would a brand campaign.

PS

Different between generic campaign and brand campaign

• Generic campaign promote a general product, we don't use brand name on it and use it in acquisition stage
• Brand campaign promote specific brand and use it in conversion stage

TV

We can look at the impact that would have on our direct brand, PPC, SEO that goes to the homepage, emails that have been sent that specifically opened through that period after launching a tv campaign

The really important thing that attribute enables you to do, which you can't do in other activities you can understand its value not just on the conversions that happen instantly but the conversions that happen in the future.

Just because someone saw a TV advert today doesn't mean they're going to buy today, But did it increase their likelihood to buy in the future

SEO

We have the full consumer journey, We know what landing pages rank for what terms.

We can now have a much clearer understanding about wherein the particular journey those types of landing pages appear most, We can then start to understand whether we want to focus more at the second or third stage in the journeyer is it more beneficial to us as an organization to focus at the end of the journey, This shows you exactly where the opportunities are for you to change consumer behavior to drive the most incremental revenue at the right point in the journey.

SEO is what impact a change has in that journey based on where it is in the journey in the first place.

Are you going after the right keywords to truly change the consumer perception of what your brand means at the point in the journey.

Display

Display is bought in a CPM model which means cost per mille, a thousand impressions.

you can't see the impact very often a click-based model, the model that is most often used is a post-impression model, So if a person views any impressions and then within a period of time visits the website and then ultimately converts, that conversion is given some attributed value back to the impression that was served.

Do you make an ROI, does it change the perception that that user has when they see the impression ?

1- we need to see every single impression, We need to see the impression at a user level, So are we able to get the data to say "User one, two, three, four" saw that particular impression?"

2-  the campaign information. So, did that person see Campaign A or Campaign B?

3- when the user sees that impression 

“ We need a timestamp, because we need to put this in the context of the user's full journey, see what impression drives conversion and what not yet.”

Affiliates 

understand the full customer journey and the value that every interaction has within that journey.

What you're able to see, is very often that the activity that happens at the end of the journey is massively overvalued in your typical last interaction attribution models.

Type of affiliates : 

• Voucher code or discount
• Influence marketing for awareness

what you can do which is really important in affiliates is understand the time difference between the last visit which is typically the converting visit and the visit before that

Emails 

You send a few 1,000 if not 100,000 emails and then those people either open them or they don't.

They either click on them or they don't and then those visitors then either buy or they don't. So how can attribution help you here?

in your attribution model ,you're going to see a multitude of other things.

You can put a pixel in emails that will tell you whether a person has opened it or not.

You can also then monitor the clicks, When that person clicks on that email and ultimately lands on your website, and you can also have campaign parameters within that email that will tell you the individual type of email and where in the email that person is engaged and then when they come to the website

You're going to be able to see the value of a send and the value of an impression, which is an email open and ultimately, that person will come to the website after clicking on the email.

Typically, the majority of emails are sent to people who already know about you in some way whether they've signed up on a lead form on the website or a newsletter form.

In the next article we will cover the rest of attribution model and dig more into the minidegree