“Half the money I spend on advertising is wasted, the trouble is I don’t know which half”,

A famous quote that holds true for many D2C brands

I have cut the jargons, explained attribution models and created a simplified solution that you can use for your brand

Read on

Scenario 1:

You are a D2C Brand, and majority of the sales are from your brand website.

Example: Let us take a hypothetical brand “Wholesome Keto Biscuits”

A customer sees an ad on Instagram, clicks and visits the website ( Click 1)

On the website, she leaves her email ID in a pop-up that promises to give her health tips to lose weight

After a week, she receives a mail and loves the content.

She decides to check out the product again( Click 2), adds to cart but for some reason doesn’t buy.

After a few days, she sees a remarketing ad offering 20% limited time discount on one of the news websites( Google Display Ads), clicks( Click 3) and completes her purchase.

So, how much credit do you give to the different sources?

How do you decide how much budget to allocate across different channels?

How do you figure out which channels are giving you the highest ROI?

Solution: Attribution Modelling

There are 4 attribution models which are commonly used:

a) Last Click b) First Click c) Linear d) Custom Weightage

a) Last Click Attribution:

This is the default attribution model that is there in Google Analytics

100% of the credit goes to the last click( Google Display in this case)

Very simple, but you understand the problem right?

Overestimating the role of one channel significantly

b) First Click Attribution:

Similar to last click, but in this case 100% of the credit goes to the first click( Instagram in this case).

Similar problem of overestimating the role of one channel

c) Linear Attribution:

Here, you split credit equally across all the 3 channels ( In this case, allocate 33.33% credit to each of Instagram, Email and Google Display)

d) Custom Weightage Model:

Basis the category, you decide weightages.

If you think getting quality audience for the first time is most difficult, allocate more to the 1st click.

If you think, getting someone to come back is most difficult, allocate more to the later stages.

Basis your product ( short/long buying cycle, impulse/evaluated purchase, price, TG etc), go ahead with the custom model and allocate weightages accordingly once you go through the Google Analytics path to purchase data for few months

After taking into account all such purchases with different journeys, you will get a consolidated channel wise percentage attribution that tells you what % of sales are being attributed to:

a) Social b) Display c) Search d) Email e) Direct f) Any other acq channel

Now, divide the channel wise attributed revenue by the channel wise costs incurred to find out which are your most efficient channels.

And accordingly scale those up.

And see your overall ROI go up significantly.

Scenario 2:

You are a D2C brand which also sells on Amazon.

The sales are divided 70-30 across Amazon and D2C.

Attribution here is much more challenging. But with a framework I have tried to simplify this. Read on

Example: We have our hypothetical brand “Wholesome Keto Biscuits”. Click 1 and 2 remains the same as scenario 1

After click 2, the customer sees time to delivery as 7 days on the site

She thinks of checking on Amazon. She goes to Amazon and searches “Wholesome Keto Biscuits”.

It shows 48 hour delivery, was on a deal pricing and she purchases it immediately.

Google Analytics never registers this as a conversion, and you would think the ads as well as the Email didn’t work

No traditional attribution models would work here. So lets get creative

a) For all sales that happen on Brand Website: Follow the same process as in Scenario 1

While it might happen that some consumers might have the reverse journey( discover on Amazon and buy from brand website), but it would be the exception and not the norm

b)For all sales that happen on Amazon- These can be attributed to either:

1) Amazon Ads 2) Organic Discovery on Platform( rare for early stage brands) 3) Brand Searches on Amazon ( like in our example)

From the Amazon Advertising Dashboard, you will find what % of sales were influenced by ads.

For the remaining %, allocate it in exactly the same ratio as you did in point a.

The key assumption is organic discovery for a new brand is negligible and any brand searches that happen on the marketplace is because of the marketing interventions

Both of these are not too far from the truth

Now combining a and b, you can find the total sales due to ads( other than platform ads) split channel wise.

Divide it by the channel wise spends, and scale up the channels which are most efficient.

This might not be 100% accurate as it has few key assumptions.

But as a brand owner at an early stage, all you need is good directional inputs on which channels of acquisition are working, at what cost, and accordingly try and scale them

And using this, you get all of these.

If you want to get deeper about marketing attribution with more complex use-cases( offline Ads, offline sales incorporated into the mix), feel free to DM.

Would love to exchange notes, ideas and current Martech solutions around this problem

If this was useful,

a) Retweet the 1st tweet in the thread

b) If you want to deep dive into different aspects of growth, attribution, e-commerce and brand-building, follow me @arndm9832 for more such threads