Undoubtedly, Google’s fastest growing product, Shopping, is one of the hardest products to master cost effectively. And it’s getting harder by the day as Google encourages more and more retailers on to it, by offering a new automated shopping solution. Where does that leave you as an agency or as an in-house practitioner? How can you protect yourself from automation and your campaigns from spiralling costs and competition? Find out how Google’s top industry trends such as increasing walled gardens, shift to transaction-based ad formats and machine learning can work to your advantage rather than against you.
SearchLeeds 2018 - Elizabeth Clark - Dream Agility - The future of Shopping
1. GL does not show for generic search terms
Galeries Lafayette opportunity – to better optimise Product Titles & Description
Elizabeth Clark
CEO
Dream Agility
글린 포우디치
최고기술책임자
Dream Agility
2. Dream Agility is a…
“ multi award winning, patent pending, 2nd generation ad tech with machine
learning and AI, a Google Premier Partner, and one of only 42 Google
Shopping Partners globally. Our software sits between a business’s website
and the search engines to optimise data and results with the equivalent
impact of an army of people – it goes beyond just bidding, and can change
the dynamics of how a business functions ”
The twitter pitch
3. We have Awards and Recognition from America to Europe, and Korea ..
@Dreamagility
4. We work with the following platforms and search engines...
@Dreamagility
5. Googles assertions today as top industry trends….
1. Increasing walled gardens
2. Shift to transaction based ad formats
3. Machine Learning : Competitive differentiator for you.
6. Current YoY % revenue growth rate
35%
Dream Agility
Top 10 SEM Partner
Agencies
159%
7.
8. Leveraging Walled Gardens to your advantage
Recommendations
• Leverage walled garden data as much as possible
• Free customer insights from walled gardens as a starting point.
• Add your own insights through your own systems
• If you don’t have one partner with an Adtech platform.
• Differentiate your distinct customer insights so your customers see the
value you provide.
• Competitive differentiation increasingly based upon customer insights,
not relationships.
10. Machine Learning the competitive differentiator for you
Machine Learning is…
A Branch of artificial intelligence.
A way to solve problems without explicitly codifying the solution.
A way to build systems that improve themselves over time.
Operated in either supervised or unsupervised modes.
12. Q. The root of the problem?
A. Manual processes beginning in merchandising.
13. Gucci Montre Femme
12 Diamantes noir
Gucci Montre Femme
noir diamantes GG2570
• Change the text in the Ad – NOT
on the website
• Test multiple Ad titles and
descriptions.
• Choose titles and descriptions
that closely match the best
performing search terms to keep
costs low and conversions high.
• Bids set per child product - not ad
group.
Machine Learning transforms the problem ….
14. • Text Ads not product specific
• Landing on a general page
• Ad extensions pointing to
femme/ladies and
Homme/men when it was a
ladies related search
• No price in the Ad!
Use machine learning on text ads too….
15. The role focus will be….
• more strategic
• Less manual
• Have closer involvement with other departments further up the food chain
• Control the source data
• Be able to predict opportunity costs.
• Lead by new metrics such as cost centres for each individual product.
• More aligned to innovation and company metrics.
Machine Learning will change the PPC managers role forever….
16. Manual structures raise cost exponentially ….
Search terms
Ad Group
multiple
products
Clicks Conversions
Pocket punch
A4 Product 1 10
1
Premium
plastic wallets Product 2 10
No punch A4
acetate wallets Product 3 10
A3 plastic
folder Product 4 10
Multiple products in an
AdGroup
Unable to tell which search
term has triggered the
conversion against which
product
Unable to identify which
redundant search terms to
remove to reduce costs
Loss making products left live
17. Where Googles structure strategy falls down ….
Account on ZERO
conversions only
savings Cost last 30
days
% Cost
Savings
Website 1 A 2,860.08 67,603.44 4.23%
Website 1 B on DA Auto PLA 6,888.47 47,246.28 14.58%
TOTAL 9,748.55 114,849.72 8.49%
18. Use Visual AI to:
Build feeds from images.
Identify errors in fields that
Google can’t spot.
Correct 404’s on Image Links.
Create feeds in multiple
languages.
Create colour libraries specific
to your customer base likely
choice.
Train the language to suit the
choices of your customers.
20. Driving the conversation rather than trying to respond to
demands on a drop in advertising efficacy.
21. Driving the conversation on pricing rather than wasting money
and driving down ROAS on products that wont sell.
22. Nobody cares about impression share and clicks, the big questions are…
“Are we making profits based on real margins & ad cost or just hitting a CPA?
How are our biggest sellers performing?
Are they in stock?
Have we forecast enough stock?
Should we be re-ordering now?
Merchandising say our best sellers are in stock, BUT how long were they
actually out for?
Are we price competitive on key products / lines?”
23. • 2.5 million products in the
feed
• Changes prices every 15
minutes or every time they
sell something .
• 600+% increase in ROAS with
Machine Learning and AI.
Music Magpie ….
24. • 25% fall in costs
• 425% increase in sales yoy
• 44% increase in ROAS with
Machine Learning and AI
My Generator….
25. Elizabeth Clark
CEO
Dream Agility
글린 포우디치
최고기술책임자
Dream Agility
https://www.linkedin.com/in/elizabeth-clark-da/
Elizabeth@DREAMAGILITY.COM
www.dreamagility.com
@dreamagility