2. • VC view on understanding customer acquisition costs
• Challenges in marketing attribution
• A quant marketers’ approach to ATL advertising
Q&A throughout please!
Agenda
3. Invest $0.5M - $15M
into tech companies
with $B potential
Invest across Europe,
help businesses go
international
Have invested in
~150 companies over
last 15 years
3
8. We are raising €15M to scale up marketing
We know our economics work
Last quarter we acquired 2000 customers at a CPA of €18.50
Average value of a customer is €25
The problem
13. 1. Better to track CPA than CPV, as conversion rate from visitor to
customer often varies dramatically by channel.
2. Don’t include brand SEM in your SEM CPA.
3. The only truly free channel is people directly typing in your URL. Other
‘free’ channels (SEO, CRM) have some variable cost.
4. Differentiating between acquisition costs of new vs. returning visitors
requires investing time and money into your web analytics system.
5. Customers which take time and multiple steps to convert are tricky.
Need to monitor both ‘Cost per Sign Up’ conversion.
6. Discounts need to be included in CAC and assigned to correct channel.
Best practice
14. Audit:
• What is my CPA by channel, what can I decrease this to?
• What can I increase my customer value to?
• Would each of my acquisition channels make sense at this CPA and value? If not,
turn off the ones that don’t. high.
From analysis to action
Growth:
• How do I grow ‘free’ channels? (top tip: start with CRM). What is a realistic
target for this?
• How far can I scale up paid channels before they increase to uneconomic levels?
• What other marketing channels should I look at? (10-20% of budget on tests)
• What growth hacks are not too hackneyed?
• Start layering in paid acquisition, starting with the cheapest channels, until you
hit your maximum budget or CPA creeps up too high.
15. This is still a simplification
Caveats
Ignores halo effects, attribution, difference in values of
customers, return rates, increase in acquisition costs as
channels get tapped out, decrease in acquisition costs as
conversion rate improves, seasonality etc etc etc
If your marketing budget goes above £10M, worth hiring an
econometrician to figure these out
But it is probably the right level of detail for a startup at
Series A stage
16. • VC view on understanding customer acquisition costs
• Challenges in marketing metrics
• A quant marketers’ approach to ATL advertising
Q&A throughout please!
Agenda
17. “The dangerous seduction of lifetime value”
http://abovethecrowd.com/2012/09/04/the-dangerous-seduction-of-the-
lifetime-value-ltv-formula/
Challenges / watch outs:
• Lifetime values beyond year 1 are speculative at best
• You don’t know what your churn rate will be. Low sample sizes fom
early cohorts, not representative
• Need to include all variable costs
• The values ‘tug at one another’: e.g. increasing price increases churn
18. Attribution
• If a user touches multiple marketing channels before becoming a
customer, which one do you attribute the sale to?
• Best approach starting out is last-click attribution. Watch outs:
• Overstates impact of affiliates and retargeting, by ~30-40%
• Understates impact of SEM and SEO by 10-20%
• Understates ATL advertising and PR
• Google Analytics is a good enough tool to start with
• As business gets more sophisticated can go for weighted first click /
last click, or weight equally through purchase cycle
• At scale worth building full attribution model. Need to do in-house,
avoid black boxes and agencies
19. Direct traffic
• Direct site visits (brand searches / app store downloads) are often a
big % of conversions
• They don’t happen by magic, but how do you attribute?
1. Make sure your analytics is set up properly and you are tracking
everything you can (e.g. clicks from emails)
2. Track referrals (e.g. with MentionMe)
3. Strip out all attributable traffic
4. Strip out repeat customers
5. Look for spikes in sales, link back to PR (or app store featuring,
promotions or brand advertising)
6. Assume a ‘halo effect’ of paid advertising, 10-20% more conversions
7. Remainder is genuine word of mouth. If it is increasing
congratulations – you are viral!
20. Direct traffic
• Direct site visits (brand searches / app store downloads) are often a
big % of conversions
• They don’t happen by magic, but how do you attribute?
1. Make sure your analytics is set up properly and you are tracking
everything you can (e.g. clicks from emails)
2. Track referrals (e.g. with MentionMe)
3. Strip out all attributable traffic
4. Strip out repeat customers
5. Look for spikes in sales, link back to PR (or app store featuring,
promotions or brand advertising)
6. Assume a ‘halo effect’ of paid advertising, 10-20% more conversions
7. Remainder is genuine word of mouth. If it is increasing,
congratulations – you are viral!
21. Conversion optimisation
• The most important thing you can do to reduce CPA
• Always be testing
• Without action conversion tends to slightly decrease
over time
• Be careful with A-B tests:
• Results may not be significant
• Positive results are not cumulative
• Need to go back and retest as business evolves
• Doesn’t take into account long term effects (e.g. ad
blocking)
http://www.qubit.com/research/most-winning-ab-test-results-are-illusory
22. Balderton portfolio: Marketing analytics tools used
22
0 5 10 15 20 25
Chartbeat
Managed in-house
RJ Metrics
Qubit
Flurry
Adobe Omniture
Optimizely
Google Analytics
Other mentions:
• Mixpanel (x4)
• Chart.io (x2)
• Kissmetrics (x2)
• Intercom
• netstat
Number of respondents who have used each approach
Rating (/5):
5
4
3
2
1
23. • VC view on understanding customer acquisition costs
• Challenges in marketing metrics
• A quant marketers’ approach to ATL advertising
Q&A throughout please!
Agenda
24. 2001 - 2012: TV ads are a waste of money
TV advertising: the VC view
2012
2013 - 2015: TV is the best marketing channel
25. Balderton portfolio: experience with ATL channels
25
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Display ads Video ads TV Print
Status:
Scaled up, significant marketing channel
Tested, positive results, yet to scale up
Will test in 2015
Not tested, don't plan to
Tested, disappointing results
Other channels used:
• Industry-specific events (x4)
• Outdoor/ Tube (x3)
• Direct Mail
• Community meetups
• Personalized merchandise
26. Two approaches to measure impact of TV advertising
• Measure immediate impact
of each TV ad slot
• Adjust upwards for:
• Delayed site visits
• Delayed conversions
• Improvement in click
through and conversion
rates due to brand
familiarity
Bottom up Top down
• Compare overall customer
and visitor numbers to
organic baseline
• Strip out trackable
marketing channels, and
adjust for seasonality
27. 8 + 8 + 8Bottom up
20:08
Ad shown
20:00 20:16 20:24
Sales
Time
28. 8 + 8 + 8Bottom up
20:08
Ad shown
20:00 20:16 20:24
Sales
Time
Impact
of TV
TVSquared and
similar tools can
help with this
29. • Compare overall customer
and visitor numbers to
organic baseline
• Strip out trackable
marketing channels, and
adjust for seasonality
Top down
31. Watch outs on TV advertising
• “Media for equity” TV is not free. Account for it at what it
would cost in the market (~20% of list price)
• Creative counts. Test it. Don’t be afraid to make it brash:
say your brand multiple times
• Track independently from your agency e.g. with TVsquared
• You can test TV with €100K, but the answer will always be
‘it doesn’t work’. Proper test more like €500K
• TV advertising doesn’t work for every company