In 2019 Adido were headline sponsors for ABTA's Advanced Social Media Trends in Travel event with members of the team delivering several talks during the course of the day.
Our CEO, Andy Headington, led the penultimate seminar of the day with insights into tracking the ROI from your social media platforms
4. GOOGLE ANALYTICS
LIMITATION 2
TRAVEL JOURNEYS
TO ONLINE
CONVERSIONS ARE
RARELY ONE
INTERACTION
LONG…
…therefore standard results are
unrepresentative of real life
9. MAKING MARKETING INVESTMENT DECISIONS
THE TIERS OF DATA ANALYSIS AND METRICS
REACH / VIEWS / FOLLOWERS
CPC / CTR
CPA / CPL / CONV. RATE
SALES / REVENUE
ROI /
PROFIT
Real time benchmark
statistics – how well are
you competing in the ad
space?
Initial indicators of success based on
‘what good needs to look like.’
Often guesstimated based on ratios but
can be calculated from online stats.
Reliant on information from business intelligence.
Could involve one-off or multiple purchases from same
customer over a period of time
Only possible by accounting for all costs and all revenue
involved in generating the sale (snapshot in time or lifetime)
10. GOING THAT NEXT LEVEL WITH YOUR BUSINESS INTELLIGENCE
CLOSING THE LOOP
11. CLOSE THE LOOP ON ONLINE CONVERSIONS
MATCH OFFLINE SALES TO ONLINE LEADS
VISIT
VISIT GET QUOTE
HOLIDAY
BOOKED
VISIT
START
ENQUIRY
HOLIDAY
BOOKED
HOLIDAY
BOOKED
GET QUOTE
START
ENQUIRY
START
ENQUIRY
QUOTE
“CONVERTED”
Offline
Online (GA)
TRAFFICSOURCES
GA
GA
13. THE IMPORTANCE OF A REFERENCE ID
PRESENT ACROSS MULTIPLE DATA SOURCES
14. TIE ALL DATA SOURCES TOGETHER
MULTIPLE SOURCES OF INFORMATION WITH CONNECTING ELEMENTS
Web analytics information
about online lead capture
• Reference ID captured in URL,
but part of a longer query
string
• Extract ID
• Report metrics available
• Date – Year, Month, Day
• Default channel
• Source
• Campaign
• Keyword
• Conversions
Master data about each
customer sale & quote initiated
• Date
• Reference ID
• Status (quote or sale)
• Product/policy purchased
• Value
• Quote reference
• Policy reference
Spend & performance data
from advertising platforms
• Spend
• Impressions
• Clicks
• CTR
• Avg. CPC
• Quotes
• Conversion rate
15. ASSESSMENT OF PERFORMANCE
ONLINE WEB TOOL
DATA
GOOGLE
ANALYTICS DATA
COMBINED WITH
OFFLINE SALES
DATA
WEEKLY UPDATE
• We assess month of conversion
vs. month of investment
• Budget delivery and benchmark
performance statistics
• Online conversions are reported
side by side with offline
conversions attributed to month of
investment
ANYTHING WHICH
CANNOT BE STITCHED
TOGETHER WITH
ONLINE TRAFFIC
• Section which identifies the
“unknowns” which can be pure
offline activity or untracked
behavior (monitor for
anomalies)
16. +42%
INCREASE IN
BUSINESS
REVENUE FROM
ONLINE
(Q1 2019)
• We were able to increase PPC marketing spend as proof of profitability was more evident
• We could also keep high positions & hold our nerve on rising CPC prices ue to inflation and competition
because we knew ad activity was still profitable
• We could explain variances in average policy values based on seasonality to manage expectations and
avoid wasted time on pursuing red herrings in data
THE IMPACT
BETTER DATA INTELLIGENCE LEADS TO BETTER DECISION MAKING
+90%
INCREASE IN
PPC POLICY
SALES &
REVENUE (YOY)
(May-19)
+4%
INCREASE IN
TOTAL
BUSINESS
REVENUE
(Q1 2019)
• We could grow new revenue from online sources (the key source of ‘new business’) and
achieve a greater share of overall business revenue.
• We matched online leads through to additional income streams (referral partner
commission) to offer additional impact of digital advertising on income streams – circa. +6.5%
extra revenue.
• We have intelligence at campaign, keyword, and policy type level so we can dial advertising
activity up or down or identify gaps to exploit with new search terms or channels on a weekly
basis if required
17. CLOSING THE LOOP
USING THE GCLID / ADID VARIABLE
When ads are clicked from Google (or
Bing) the URL typically contains a
string of letters/numbers on the end
unique to the click (the Google CLick
ID)
Alternatively, the FB ADID could
captured at the enquiry stage and then
pushed through to your database for
matchback down the line
This information can be extracted and stored in your
database to be re-uploaded into Google Ads at a later
date
Typically, used for adding sales data (converted at a later
date) from a lead triggered online from your digital ads.
gclid gets saved
against lead
record in database
When lead
becomes sale this
record’s gclid is
extracted
gclid record gets
imported into
Google Ads
Stats added to
Google Ads
reports
Optimise against
this metric
18. THE IMPACT
• This client is in a saturated, stagnant market and yet
in the last 3 years we have managed to improve ROI
from 20% to 24% across PPC & social by applying
this layer of data analysis to optimization decisions
• We were able to push highly profitable terms and
not be beholden to CPA targets as stringently
• We could make decisions on which terms were
losing money despite having low CPAs and good
lead volume and turn them off
• We could initiate automated bidding rules based
on these additional metrics rather than CPA which is
less of an indicator of success than profitability
+9%
INCREASE ROI
from PPC YoY
(Jan-Aug 19)
+13%
INCREASE IN
PROFIT FROM
PPC YOY
(Jan-Aug-19)
20. LEADS AND SALES DON’T OCCUR IN A VACUUM
TYPICALLY PATHS TO CONVERSION ARE MORE THAN ONE INTERACTION IN LENGTH
BUT HOW
DO YOU
APPORTION
CREDIT?
Factors in
time lag
Good for
low volume,
high value
conversions
21. GOOGLE ANALYTICS
MODEL COMPARISON TOOL EVALUATE
THE IMPACT
BASED ON
DIFFERENT
RULES
At this stage it will be an assessment of what feels right.
Or, if you decide to focus on one model and make decisions using its insight, what
happens to your overall performance?
22. HUBSPOT / CRM
SINGLE CUSTOMER VIEW
VALUE OF
DEAL
CAPTURED
CONTACT
ASSOCIATED
WITH DEAL
ATTACHED &
ACTION
HISTORY
AVAILABLE
Manually trawl through
interaction history and
identify/count how
many interaction points
involved in sale
Determine your
attribution rule
Deal 1 Deal 2 Deal 3
1 Organic Email Email
2 Direct Direct Email
3 Email Email
4 Email Email
5 Email
6 Email
7 Email
8 Email
9
10
Interactions 4 2 8
Apportion credit (equal)
Organic 0.25 0 0 0.25
Direct 0.25 0.5 0 0.75
Email 0.5 0.5 1 2
Apportion credit (last)
Organic 0 0 0 0
Direct 0 1 0 1
Email 1 0 1 2
23. • Able to assess the impact lead generation by source is having on final
sales numbers in order to dial up / down the activity or cease
altogether
• Makes it possible to compare traffic sources and marketing
investment and understand lag time to convert by interaction events
• It is however a time-consuming process, especially when you need to
back date / reconcile performance from previous months every month
THE IMPACT
BETTER
reflection of
what is actually
happening
24. • By matching booking month to arrival date, we were able to predict in advance when capacity was being sold
• We were able to break this down by park to see which parks / products needed more attention (budget) than others to help with the overall target
• We used data to help plan budgets & increase revenue and profit to help battle loss of inventory & fend off competition
THE IMPACT WE MANAGED
TO INCREASE
BOOKINGS BY
£1 MILLION!
26. YOU’LL NEED A TOOL TO SUPPORT YOU
COMBAT TIME CONSUMING MANUAL TASKS WITH A 3RD PARTY TOOL
Other 3rd
party tools
exist – do
your
research!
Integration
with CRM
Indefinte
look-back
window
27. SET-UP
• Ensure you have an ID to match all (relevant) data sources together
• Identify the marketing / performance metrics which will inform the best decision making and work within the
constraints of their accessibility (regularity / origin of data)
• Should you / can you speed up decision making by applying metrics higher up the chain using ratios?
• Update these ratios overtime as more business data becomes available
• Look beyond last click to get a universal view of your marketing efforts for both online journeys and full
customer journeys – determine which rule fits your business best
TO CONCLUDE
GOING BEYOND LAST CLICK + ONLINE
28. REPORTING
• Get hands on with your data and manually stitch together your data sources to get a sense of what is happening
• Scope out a reporting process which can be fully or partially automated based on resources
• Seek to invest in a 3rd party tool to help automate the online > full customer journey at scale when a manual
solution becomes unmanageable or could open up new avenues for growth
TO CONCLUDE
GOING BEYOND LAST CLICK + ONLINE
29. INSIGHTS & ACTIONS
• Evaluate performance against month of investment
• Reconcile performance on a regular basis to get a timely assessment of what is often a lengthy buying cycle
• Determine the role and effect of each marketing activity to inform investment decisions and campaign
effectiveness
• Identify growth opportunities in underused marketing channels
• Reduce budget wastage
TO CONCLUDE
GOING BEYOND LAST CLICK + ONLINE
Google Analytics is the ‘go to’ analytics tool for measuring website performance
Google Analytics is the ‘go to’ analytics tool for measuring website performance
So we’ve ‘closed the loop’ on online leads (in this instance quotes) with offline sales for Event Insurance Services.
Closing the loop means feeding back sales data / closing off the conversion journey once someone has become a customer and assigning back to the source of the lead (which converted into a sale)
They typically have three types of converting journey once someone is ready to request a quote
Start quote, complete quote and buy quote online – all in one session
Start quote, complete quote but then stop their online journey – they can either save their quote which will be emailed, or phone up later to buy the policy
Start quote, but then abandon the process – if they decide to phone later, often the sale representative is able to retrieve their incomplete quote, update and convert it for them
QueryClick
Widely-used solutions including Google’s Attribution 360 (formerly Adometry) and Adobe Analytics provide only historical attribution insights that are frequently incorrect due to data cleansing challenges, preventing proactive spend adjustment.
To provide accurate, forecasted attribution insight, QueryClick has combined pioneering data science and machine-learning techniques to develop Corvidae.
Corvidae’s random forest and Markov chain machine-learning mechanisms simulate billions of future outcomes based on your historic marketing data.