Marketing Response Understanding the complete marketing ROI for PC City
Google Confidential and Proprietary
1
Background PC City is an chain of computer superstores and e-commerce operation that operates across Europe. As part of the the DSGi group PC City (& PC World) it has stores and e-commerce operations across UK, Spain, Portugal, Italy, Greece and Sweden. The business in Spain started via the acquisition of Ei System and today operates 32 stores across the country. The business started its e-commerce operation over 5 years ago and has used both online and offline media to drive traffic to its site PcCity.es. Online advertising historically has been deployed at low levels with Catalogues and TV having been the dominant historical media (~95% of PC city media expenditure).
Google Confidential and Proprietary
2
Research Objectives and Methodology Objectives • To understand and quantify the effectiveness of PC City’s marketing and media spend in driving business outcomes • This across both online and offline businesses • To understand the role of Pc City’s paid and owned media (website) in driving sales
• To optimise the marketing and media mix to generate greater revenue from the same level of investment
Methodology • Using Pc City’s internal and proprietary data (May 2007 – Feb 2010) • • • •
Sales volumes and revenues (split by channel (online vs offline) Pricing Website data (visitors, conversions) Marketing and Media costs
And data from a variety of other internal & external sources (including search volume data from Google.es) Multiple econometric models (multiplicative logarithmic time series) were constructed to understand total Marketing ROI by media channel and inform future optimal media resource allocation 3
Google Confidential and Proprietary
Understanding the PC City business Pc City owns and operates 32 stores across Spain Although only Laptops, Netbooks and Desktops were analyzed as part of this work these 3 lines of business represent approx 85% of Pc City revenue
Business Split Online/offline - Revenue
Online sales 3%
1% Reserved Online, collected Offline
96% Offline sales
Business Split Online/offline - Units
Online sales 4%
2% Reserved Online, collected Offline
94% Offline sales Google Confidential and Proprietary
4
The Offline World What are the drivers of offline sales?
Google Confidential and Proprietary
5
Understanding the drivers of offline sales 11% of Pc City’s offline in-store sales comes as a result of visits to it’s website www.pccity.es. 10000
11%
8000
Radio & OOH 1%
7000
Other drivers 1%
of offline
conversions driven by Pc City web traffic
6000
Seasonality 8%
Catalogues 15% Price promotion 2%
5000
Regular price 12%
Press 3%
4000
TV Advertising 5%
3000
2000 1000
Base Sales 42%
Google Confidential and Proprietary
06-Feb-10
06-Jan-10
06-Dec-09
06-Nov-09
06-Oct-09
06-Sep-09
06-Aug-09
06-Jul-09
06-Jun-09
06-May-09
06-Apr-09
06-Mar-09
06-Feb-09
06-Jan-09
06-Dec-08
06-Nov-08
06-Oct-08
06-Sep-08
06-Aug-08
06-Jul-08
06-Jun-08
06-May-08
06-Apr-08
06-Mar-08
06-Feb-08
06-Jan-08
06-Dec-07
06-Nov-07
06-Oct-07
06-Sep-07
06-Aug-07
06-Jul-07
06-Jun-07
0 06-May-07
Sales (units sold)
9000
Offline Sales volumes
6
Understanding the drivers of offline sales The Website is responsible for 11% of offline revenue – which is nearly 1/3 of all sales driven by marketing 11%
62% of PC City sales are non marketing drivers
15%
3%
1%
1%
5% Offline Sales volumes
12%
8%
Catalogues and the internet are the biggest drivers of offline sales.
42%
Base sales volume includes influence from existing store network, employee workforce, existing infrastructure distribution, seasonality, underlying consumer demand etc.. Base Sales
Seasonality
2%
The Website is responsible for 11% of offline revenue – which is nearly 1/3 of all sales driven by marketing
38% of PcCity sales are influenced by Marketing
Regular price
Price promotion
TV Advertising
Catalogues
Press
Radio & OOH
Other
Web traffic to pccity.es
Google Confidential and Proprietary
7
The Online World What are the drivers of online sales?
Google Confidential and Proprietary
8
Understanding the drivers of incremental online sales Catalogues 4%
Incremental online sales are driven by predominantly by price and underlying demand.
Press 3% 10%
TV Advertising 3%
500 450
Graphic quantifies the drivers of incremental sales volumes – i.e. sales above the modelled norm
400
Base 32%
Price promotion 23% Web traffic to pccity.es 25% Seasonality 10%
300 250 200
Press 3%
TV Advertising 3%
150 100 50 0
Base 32% 2007-26 2007-29 2007-32 2007-35 2007-38 2007-41 2007-44 2007-47 2007-50 2008-1 2008-4 2008-7 2008-10 2008-13 2008-16 2008-19 2008-22 2008-25 2008-28 2008-31 2008-34 2008-37 2008-40 2008-43 2008-46 2008-49 2008-52 2009-3 2009-6 2009-9 2009-12 2009-15 2009-18 2009-21 2009-24 2009-27 2009-30 2009-33 2009-36 2009-39 2009-42 2009-45 2009-48 2009-51 2010-2 2010-5
Sales (units sold)
350
Seasonality 10%
Google Confidential and Proprietary
9
Understanding the drivers of total sales 16% of the PC City Business is driven by the Web. 10% coming via research online but purchase offline, a further 4% via ecommerce and 2% reserved online.
16% 10% Via Online to Store
business driven by online
6% Via Online
Google Confidential and Proprietary
10
Marketing ROI How to optimise media spend to return
Google Confidential and Proprietary
11
Understanding the effectiveness of marketing spend Due to the natural variance in both spend and sales levels over the analysis period (May 2007 – Feb 2010) the models are able to plot response curves for each of the different media cannels
Media X
Weekly Sales (Units)
4,800
4,700 4,600 4,500
From understanding historical sales response to media spend it is possible to understand the ROI of each channel Additionally understanding the responsiveness of each media in driving sales it is possible to define an optimal marketing mix
4,400 4,300 4,200 4,100 -
50,000.00 100,000.00 150,000.00 200,000.00 250,000.00 300,000.00 350,000.00
Weekly Media Spend €
Google Confidential and Proprietary
12
Understanding the effectiveness of marketing spend An Optimal mix for PC City would mean reducing investments in TV and upweighting investment in other media channels.
Weekly Sales (Units)
Weekly Sales (Units)
4,800 4,700 4,600 4,500
Optimal spend
4,400 4,300
Historical spend
4,800
Historical spend
Increase spend in
4,200
4,700
Optimal spend
4,600
Increase spend in
4,500
Paid search
catalogues 4,400
4,100
-
Absolute values not shown due to client confidentiality – ticks represent and indicator of ‘000
Absolute values not shown due to client confidentiality – 30,000 ticks represent and indicator of ‘000 10,000 20,000 40,000
-
50,000 100,000 150,000 200,000 250,000 300,000 350,000
Weekly Media Spend €
Weekly Media Spend €
50,000
Weekly Sales (Units)
4,700
Historical spend
4,600 4,500 4,400
Optimal spend
4,300
Decrease spend in
Weekly Sales (Units)
4,800 4,800
Historical spend
4,700 4,600
Optimal spend
4,500
Increase spend in
4,400
4,200 4,100
TV not shown due to client confidentiality – ticks represent and indicator of ‘0,000 -Absolute values 100,000 200,000 300,000 400,000 500,000 600,000
Weekly Media Spend €
Press 4,300 -
Absolute values not shown due to client confidentiality – ticks represent and indicator of ‘000
10,000
20,000
30,000
40,000
50,000
Weekly Media Spend €
60,000
Google Confidential and Proprietary Note that the optimal recommendation takes into account the natural constraints (i.e. the available inventory) of each media.
70,000
13
Using the optimal media mix With the same budget a more optimal mix delivers an 6% revenue Optimal Mix for 24th Aug – 12 Oct 2010 (8 weeks)
Last year 23 Aug – 11 Oct 2009 (8 weeks) Paid Search 1%
Paid Search 9%
Catalogue 32%
TV 64%
TV 43%
Outdoor 1% Press 2%
Press 3% Outdoor 3%
Aggregate Aggregate Marketing €1.10 Marketing ROI of ROI of display within the media mix was compromised by a lack of data ROI Understanding relating to Display investments over the period analysed.
€1.40
Catalogues 42%
6% increase in revenue
Google Confidential and Proprietary
14
Total Campaign Marketing Response Curves The optimal media mix represents a 6% increase in sales
35,000
Opportunity = 6% 33,000
Optimal Mix
31,000
Historical Mix
Units (Sales)
29,000 27,000 25,000 23,000 21,000
Historical BTS Investment 19,000 17,000
15,000
Total Campaign Investment
Understanding of ROI of display within the media mix was compromised by a lack of data relating to Display investments over the period analysed.
Google Confidential and Proprietary
15
Return on investment per media
The ROI of each media is based upon the media investment and return .
Paid search is the most cost efficient media in the mix, more than 10x the closest competing media
€31.80 Online ROI €4.68
1€ invested in paid search delivers €4.68 in revenue online and €27.14 in revenue offline. Offline ROI €27.14 Some very low level cross channel ROI evident across press & catalogues.
€0.7
€1.28
€1.12
€0.94
€1.36
Understanding of ROI of display within the media mix was compromised by a lack of data relating to Display investments over the period analysed.
€1.50
Google Confidential and Proprietary
16
Understanding the effectiveness of marketing spend
Assuming all other media investments are held at the current level, it is possible to understand the effectiveness of incremental media spend on sales
Increases in Paid search investment will deliver the greatest return Further €10k in
Delivers … additional Units sold
TV
9
Press
67
Catalogues
13
Paid Search
131
Understanding of ROI of display within the media mix was compromised by a lack of data relating to Display investments over the period analysed.
Google Confidential and Proprietary
17
Understanding Digital
Google Confidential and Proprietary
18
Paid search position and revenue Increasing position of Paid search has a noticeable effect on both online and offline revenue
Average position
% increase in % increase in online sales offline sales
Generic search terms across Laptops, Notebooks and Desktop
1.0
2%
9%
1.7
1.7%
7%
3.5
-
-
Historical Paid search metrics for Averagecity Position PC Average Cost Average Position per Click
Brand Keywords
Generic Keywords
3.1
3.5
Google Confidential and Proprietary
19
Drivers of Branded search queries In addition to driving web traffic and ultimately sales, Generic Paid Search impressions also drive 4% of branded query volume
Catalogues 2% Paid Search Impressions 4%
TV 1%
Press <1%
Seasonality 15%
Base Online Queries 78%
Google Confidential and Proprietary
20
Conclusions
Google Confidential and Proprietary
21
Summary and conclusions
1
16% of PC City’s business is driven by the web
2
Each €uro invested in Paid search delivers €27.48 in revenue offline and €4.68 online
3
By shifting the media mix Pc City could increase total revenue by 6% with the same media expenditure
4
By changing average page position of paid search would increase online sales by 2% and offline sales by 9%
Google Confidential and Proprietary
22
Testimonials
Google Confidential and Proprietary
23
Testimonials
Jorge Benlloch – CEO PC City Spain We all know that the future is online, but we didn’t know until now quite how powerful it was. Thanks to this research we now understand the importance of each media channel. We know that each media plays a distinct role in the sales process but we are now able to optimise media budget to align to our business goals. We now have the data and analytics to test significantly increased spending in paid search and are already doing so.
Google Confidential and Proprietary
24
Appendix Some other interesting bits and pieces
Google Confidential and Proprietary
25
20-Feb-10
20-Jan-10
20-Dec-09
20-Nov-09
20-Oct-09
20-Sep-09
20-Aug-09
5000
20-Jul-09
20-Jun-09
20-May-09
20-Apr-09
20-Mar-09
20-Feb-09
20-Jan-09
20-Dec-08
20-Nov-08
20-Oct-08
20-Sep-08
98%
20-Aug-08
Model explains of
20-Jul-08
20-Jun-08
20-May-08
20-Apr-08
20-Mar-08
20-Feb-08
20-Jan-08
20-Dec-07
20-Nov-07
20-Oct-07
20-Sep-07
20-Aug-07
20-Jul-07
6000
20-Jun-07
20-May-07
Laptop Sales (Unit)
Model Accuracy
variation in Modeled
Actual
4000
3000
2000
1000
0
Google Confidential and Proprietary
26
Understanding the PC City business Legend Business Outcome
Online worldOffline world
+ Driver
Traditional Media
Online Search
- Driver
Store Refurbishment Holiday
SEM Online (Paid Search) Email Display
Web traffic
Offline Sales
Online Sales
Competiti ve Media Price, Base Sales & Seasonality
Width of line indicates size of driver
Short term Promo days
Sporting Rainf Events all 27 Google Confidential and Proprietary
27
About MarketShare Partners World-Class Marketing Science + Real-World Industry Expertise
All data analysis and econometric modelling was conducted by
• Strategic analytics firm deploying industry’s most advanced analytic models for marketing effectiveness, ROI and allocation • Reputation of being the most innovative firm in the category, with focus on advanced math, thought leadership and action • Significant experience with new media, including search, social networks and display • Heavy emphasis on application of analytics to make better decisions going forward • An independently owned company delivering 100% objectivity in assessing marketing effectiveness.
Google Confidential and Proprietary
28
Thanks!
Google Confidential and Proprietary
29