Reeport Partner presentation - Mixing site- and ad- centric data despite the media jungle

Mixing site- & ad-centric data
despite the media jungle
5 lessons learned from 8,000 dashboards

Your host
Etienne GAUTHERON
etienne.gautheron@reeport.io
Director, Product & Operations

We help organizations democratize data

● Spin-off of Uptilab
● Based in Paris
● Created in 2010
● 300+ employees
● Data start up of the year
@Vivatech
● Innovation in
partnership with
France’s Ministère de la
Recherche &
Polytechnique
● +€10M raised

80% time spent on data collection and formatting 20% for analyses
Data collection and preparation The last mile of data
Going from data to informed decisions

#1
Set up your reporting before
launching the campaign

What I’ve learned from Optimizely
At Optimizely, our goal was to help companies optimize something through AB testing!
The biggest obstacle to running a successful AB testing program was the lack of hypothesis.

What I’ve learned from Optimizely

What I’ve learned from Reeport
At Reeport, our goal is to help companies understand their digital investments performance.
The biggest obstacle to understanding performance is the lack of prior KPI setup.

No need to start from scratch
You’ve categorized your campaigns already?
Then categorize your KPIs

Defining KPIs
Relevant Understood Reliable Actionable Available

Setting up your reporting beforehand
I will launch a _________________ campaign
because my business goal is to _________________
as measured by _________________ and _________________.(primary indicator) (secondary indicator)

Setting up your reporting beforehand
I will launch a branding campaign
because my business goal is to make myself known to a new market
as measured by impressions # and funnel step 1 entry %.

Your next step

Your next step
And turn them into analysis templates

#1 Set up your reporting before launching the campaign
Overcoming the
online ad jungle

#2
Question the KPIs
you are given

A KPI is a proxy. It never reflects entirely the reality.
A KPI is a bias. Someone has to set it to reflect one reality (their own).
A KPI is a referee. It tells you what to do.
Wait...so what’s a KPI again?

A KPI is a proxy. It never reflects entirely the reality.
A KPI is a bias. Someone has to set it to reflect one reality (their own).
A KPI is a referee. It tells you what to do.
Understand it or create it, don’t just follow it.
Wait...so what’s a KPI again?

Google’s last click attribution model

Google’s last click attribution model
“Gives all credit for the conversion to the last-clicked ad and corresponding keyword.”

“We had an understanding that it was
digital advertising – desktop and mobile
– that was driving those sales and as a
consequence we were over-investing in
that area.”
Simon Peele
Global Media Director, Adidas

To understand their business model is to understand the bias coming out
of their reporting, numbers, and analyses.
The goal here is not to question their business model but to use it as a
key for reading their recommendations.
There’s a business model behind every KPI

#2 Question the KPIs you are given
Overcoming the
online ad jungle

#3
The relationship
with your partners is key

The promise
In the world of digital,
you have tons of data
readily available in
real-time to measure
everything.
The reality
In the world of digital,
you have too much
data thrown at you but
still lack the one you
really need.

Ask for minimum viable data
No data transparency? No good.
Relevant Understood Reliable Actionable Available

The source of most problems? Nomenclature.
[channel]_[geo]_[ad type]_[brand]_[product category]_[promotion]
[geo]_[ad type]_[brand]_[promotion]
[channel]_[geo]_[ad type]_[brand]_[product category]

About naming conventions
Laziness will cost you money.
Find one and be consistent. Really.
Be uncompromising.

#3 The relationship with your partners is key
Overcoming the
online ad jungle

#4
APIs aren’t king yet,
long live the CSV!

There are great APIs
out there

Plug-and-play API
connectors
Automated data flows
WYSIWYG editor
User-friendly viewer
interface
Design guidelines
Multiformat sharing (email,
dynamic online dashboard,
datawall, PPT, PDF, XLS,
CSV)
Alerts
Adoption monitoring
Google Cloud Dataprep
Google BigQuery
SQL
Data flow setup
Data flow maintenance
Data sanity check
Google Cloud project
setup
Data cleaning
Data storage
Data reconciliation
Dashboard UI & UX
Dashboard wireframing
Dashboard creation
Communication plan
SSO setup
LDAP setup
Enterprise-level SECURITY & scalable INFRASTRUCTURE
DATA
VISUALIZATION
DATA
PREPARATION
DATA
COLLECTION
DATA
ADOPTION

Be disciplined upfront and rely on more than APIs
Tracking plan
Data warehouse
Reporting

#4 APIs aren’t king yet, long live the CSV!
Overcoming the
online ad jungle

The source of most problems? Naming convention.
[channel]_[geo]_[ad type]_[video duration]_[product]
5 dimensions fed automatically in Reeport:
1. Channel
2. Geo
3. Ad type
4. Video duration
5. Product

AT Internet data
here
Automating all you can with Reeport

AT Internet data
here
Google Ads data
there

AT Internet data
here
Consolidated
media spend here
Google Ads data
there

From pull to push at Public Sénat

Pull Push at Public Sénat
+55% taux d’ouverture

#4 APIs aren’t king yet, long live the CSV!
#5 Automate all you can
Overcoming the
online ad jungle

#oneMoreThing
Done
is better than perfect

Questions?
Etienne GAUTHERON
etienne.gautheron@reeport.io
Director, Product & Operations

Reeport Partner presentation - Mixing site- and ad- centric data despite the media jungle

More Related Content

What's hot

Similar to Reeport Partner presentation - Mixing site- and ad- centric data despite the media jungle

More from AT Internet

Recently uploaded

Reeport Partner presentation - Mixing site- and ad- centric data despite the media jungle