Lessons learned from
applying PyData to our
marketing organization
PyData + Marketing = ???
Jose Luis Lopez Pino @jllopezpino
First some bragging!
- We have grown x3 our marketing efforts.
- We have reduced ~90% of the time spent on
creating a new ad.
- Launched in 7 new markets.
- We haven’t grown the team.
- No expensive marketing software.
Two advices before starting...
You won’t go far without domain expertise
Two advices before starting...
Marketing is a fast-paced competition
What is GYG?
Marketing Technology
Create, control and target online ads
Create
- Ads that are relevant to the user.
- Within the limitations of the advertising
service.
- Send them to the best landing page.
Control
How much should we bid?
How much we expect to receive from this ad?
Ranking factors?
How good are we doing?
How much room for improvement do we have?
Target
- Keywords, keywords, keywords!
- Remarketing lists.
- Reach relevant audiences.
- Language.
Three steps
- Set up the infrastructure and tools
- Data retrieval and storage
- Analyse, automate and train models!
Infrastructure and tools
- Dedicate server?
- Schedule requests?
- Queue requests?
- Multiple nodes?
- Storage? Databases? Structured?
- Talk to other systems?
- Do you need an interface?
- How to deploy?
Data retrieval and storage
- Data modeling?
- Alembic (database migration tool) to make
changes in your DB and keep track of them.
- API migrations and sunsets: write better
code!
- Rely on third-party systems?
- Optimize for speed?
Data retrieval and storage
- Scrape without being banned?
- Extract data from other systems of the
organisation.
- Users also input data.
Analyse and automate
● When to use SQL? When to use pandas?
○ Complex pivot tables with SQL?
○ Load all my keyword space again and again in
pandas?
● Schedule queries to provide data for
spreadsheets every day.
● Allow marketers to make changes that they
can’t do with any other interface.
Some examples
- What are the products that people show
interest at this time of the year?
- Customer segmentation. What are the best
attributes to segment them?
- How do I estimate the potential size of a
market that I don’t know?
- Monitor important ranking changes.
- What are my competitors doing?
Some examples
- What are the outliers in our accounts that
need human attention?
- What are the most important keywords for a
particular page?
- What are the products that I need in my
marketplace?
Some ML examples
- Are we going to sell out this product?
- Sentiment analysis on customer reviews.
- Regression model of our ROAS.
- How to cluster our adgroups to make
decisions on them?
And we are still learning
Data scientists
…
13.03.2015 GetYourGuide AG 21
statistics
forecasting and estimation
pandas, scipy, R
…
13.03.2015 GetYourGuide AG 22
Engineers
databases & data mining
machine learning
python
visitor tracking & metrics
How to measure success?
Takeaways from this talk
- Marketing is not the latest buzzword, but it’s
fun to do for data-driven people.
- Technology can have an enormous impact
on the marketing results.
- And the PyData stack provides tools to do it.
That’s all!
Jose Luis Lopez Pino
@jllopezpino

Lessons learnt from applying PyData to GetYourGuide marketing

  • 1.
    Lessons learned from applyingPyData to our marketing organization PyData + Marketing = ??? Jose Luis Lopez Pino @jllopezpino
  • 2.
    First some bragging! -We have grown x3 our marketing efforts. - We have reduced ~90% of the time spent on creating a new ad. - Launched in 7 new markets. - We haven’t grown the team. - No expensive marketing software.
  • 3.
    Two advices beforestarting... You won’t go far without domain expertise
  • 4.
    Two advices beforestarting... Marketing is a fast-paced competition
  • 5.
  • 6.
  • 7.
    Create - Ads thatare relevant to the user. - Within the limitations of the advertising service. - Send them to the best landing page.
  • 8.
    Control How much shouldwe bid? How much we expect to receive from this ad? Ranking factors? How good are we doing? How much room for improvement do we have?
  • 9.
    Target - Keywords, keywords,keywords! - Remarketing lists. - Reach relevant audiences. - Language.
  • 10.
    Three steps - Setup the infrastructure and tools - Data retrieval and storage - Analyse, automate and train models!
  • 11.
    Infrastructure and tools -Dedicate server? - Schedule requests? - Queue requests? - Multiple nodes? - Storage? Databases? Structured? - Talk to other systems? - Do you need an interface? - How to deploy?
  • 12.
    Data retrieval andstorage - Data modeling? - Alembic (database migration tool) to make changes in your DB and keep track of them. - API migrations and sunsets: write better code! - Rely on third-party systems? - Optimize for speed?
  • 13.
    Data retrieval andstorage - Scrape without being banned? - Extract data from other systems of the organisation. - Users also input data.
  • 14.
    Analyse and automate ●When to use SQL? When to use pandas? ○ Complex pivot tables with SQL? ○ Load all my keyword space again and again in pandas? ● Schedule queries to provide data for spreadsheets every day. ● Allow marketers to make changes that they can’t do with any other interface.
  • 15.
    Some examples - Whatare the products that people show interest at this time of the year? - Customer segmentation. What are the best attributes to segment them? - How do I estimate the potential size of a market that I don’t know? - Monitor important ranking changes. - What are my competitors doing?
  • 16.
    Some examples - Whatare the outliers in our accounts that need human attention? - What are the most important keywords for a particular page? - What are the products that I need in my marketplace?
  • 17.
    Some ML examples -Are we going to sell out this product? - Sentiment analysis on customer reviews. - Regression model of our ROAS. - How to cluster our adgroups to make decisions on them?
  • 20.
    And we arestill learning
  • 21.
    Data scientists … 13.03.2015 GetYourGuideAG 21 statistics forecasting and estimation pandas, scipy, R
  • 22.
    … 13.03.2015 GetYourGuide AG22 Engineers databases & data mining machine learning python visitor tracking & metrics
  • 23.
  • 24.
    Takeaways from thistalk - Marketing is not the latest buzzword, but it’s fun to do for data-driven people. - Technology can have an enormous impact on the marketing results. - And the PyData stack provides tools to do it.
  • 25.
    That’s all! Jose LuisLopez Pino @jllopezpino