guillaume “g” cabane...- facebook campaigns are optimized towards “website conversions” but...
TRANSCRIPT
Crushing competition
@guillaumecabane
Guillaume Cabane, Head of Growth Marketing, Mention
Guillaume “G” CabaneCEO @ GROWTH EX MACHINA
50 -> 200
4 growth teams and counting
50 -> 300
15 -> ?
15 -> 30
Now I am a Growth Advisor
WEB TECHNOGRAHICS
TOPIC SURGING
CUSTOM SITE SCRAPING
Chat
Web content
Advertising
Account, User and Campaign Record
Physical goods
Output Channels
VISITED WEBSITE
SCORE PROSPECTING
COLLECTION ACTION
Intent capture
Enrich & Fit scoring
SaaS reviews
Visited Website
Installed Web Tech
Reading About “Livechat”
Hiring Sales Director
TIME
Demo Request!
Predicting the lead value based on intent
SCORE: 0.12Intent
SCORE: 0.33Intent
SCORE: 0.61Intent
SCORE: 0.88Intent
VALUE: $130Prediction
VALUE: $200Prediction
VALUE: $350Prediction
VALUE: $600Prediction
@guillaumecabane
What makes a company succeed?
A distribution moat
@guillaumecabane
What’s the role of Marketing?
@guillaumecabane
Do we agree?
1: a better experience increases conversion2: with more budget, we can create a better
experience
@guillaumecabane
The challenge: We don’t have more budget *per user*
Channel Saturation & Rising CAC
@guillaumecabane
Distribution across every major channel has become more expensive
-25% CTR on Page 1 SEO in 2 years. -55% for mobile
@guillaumecabane
Inventory is stable … so costs go up
Facebook CPC's increased 136% just in 2017
CAC has grown >50% in the past 5 years
You can’t bruteforce your way into a market through sheer spend
Source: ProfitWell
CAC has
increased by 70% in the past 5 yearsbut LTV has remained constant.This cannot go on anymore.
The pitfall : averages
Your KPIs, as a uniform and representative understanding of each user at that stage, is fundamentally wrong
Huge variance of values per user you want to separate those based on their future value
Sale
s Ve
loci
ty in
$/m
onth
Company size5 50 200 1,000 10,000
Sales Velocity
Cost of sales
Contract size5 50 200 1,000 10,000
Enterprise
ABM
Self-Service
Marketing Automation
Likelihood to buy
Competitors budget per lead
Our budget per lead
high right better budget = better experiencelow left => bid is not profitable, let competitors loose money
Likelihood to buy
Competitors budget per lead
Our budget per lead
high right better budget = better experiencelow left => bid is not profitable, let competitors loose money
@guillaumecabane
Because we are bidding the right amount ...
we can spend more on the right people when our competitors can’t.
We won’t spam with emails or ads people who aren’t likely to buy our products.
@guillaumecabane
There is no insane innovation going on here.
All of this what derived from B2C bid optimization / retargeting
Building a moat to crush competition
@guillaumecabane
How can we built delightful experiences?
Ŵŵ
:KDW�LV�WKH�PRVW�LQIOXHQWLDO�LQ�\RXU�SXUFKDVH�GHFLVLRQ�SURFHVV"
:Ŵ��Ĩī�ŴŲųŻ�UǛųŵŸŴ
highlight the two bottom categories. Write “buyers don’t want us”
³:H�XVXDOO\�HQJDJH�D�VDOHV�SURIHVVLRQDO�RQO\�ZKHQ�ZH�KDYH�PDGH�D�SXUFKDVH�GHFLVLRQ�´
:Ŵ��Ĩī�ŴŲųŻ�UǛųŵŸŴŴŷ
����
@guillaumecabane
Creating a better experience … in outbound
@guillaumecabane
Good email G ….We’re missing an
illustration
@guillaumecabane
in Chat
Display LivechatFirmographic ScoreGet IP Address Returns DomainCOMPANY OBJECT IF SCORE > 0.9
Deliver coffeeBot data collection Create order
@guillaumecabane
in web pages
Change HomePage
Firmographic ScoreGet IP Address Returns DomainCOMPANY OBJECT
Intellimize Confidential
Optimizing for multiple weighted goals
Traditional approach With AI
One goal Multiple weighted goals
$100 $150 $200
$325 $250 $175
Note: All values are illustrative
$X
@guillaumecabane
What about Paid?
© 2019 MadKudu Inc. Proprietary and Confidential
The Challenge
- Facebook campaigns are optimized towards “website conversions” but this does not correlate to value
- Commonly the quality of facebook leads is low (<20% of good fits) and tends to deteriorate over time because of the aforementioned optimization criteria
- As a consequence CPAs increase and the CAC can get out of hand
Who you pay for
Who you’d like
https://andrewhubbard.co
Likely to Buy
© 2019 MadKudu Inc. Proprietary and Confidential
MadKudu predicts the value of each visitor and tells Facebook to bid only on your best leads
Your ad
© 2019 MadKudu Inc. Proprietary and Confidential
MadKudu predicts the value of each visitor and tells Facebook to bid only on your best leads
Your ad
MadKudu predicted values are sent back to facebook to train the AI to perform the best bidding
Facebook learns from that lead value to optimize spend moving forward for the rest
of the audience
MadKudu predicts the lead value in real time
New lead fills a form with their
© 2019 MadKudu Inc. Proprietary and Confidential
The results
Strategy Cost per Lead % Qualified Leads Expected Return on Investment
Expected Payback Period
Optimize to All Leads Baseline
Optimize to Good Leads Only
-27% 165% +200% -66%
Optimize to Value -49% 247% +766% -88%
Leveraging user revenue predictions in your day-to day
source: http://christophjanz.blogspot.com/2016/06/a-better-way-to-visualize-pipeline.html
rewrite name of stages , or add a marketing stage on topfallacy of averages
magasinstock en rayn-> chaque produit as-til la meme valeur? Non!
-> valeur entrant, ou du stock, doit etre modulee en fonction de la valeur de chaque produit … et de sa capacite a etre vendu
Learnings from Building Growth teams
@guillaumecabane
Guillaume “G” CabaneGROWTH ADVISOR
Scan this for Linkedin