Chapter 4 Empirical Analysis
4.4 Summary
According to the results, some attributes such as “foundation” and “open source”
have strong impacts throughout most of the stages. These kinds of keywords are worth buying and keeping exploring. Some attributes are only influential in particular stages.
For instance, the computer science related event has tendency to accrue clicks, while student camp event has impact on bounce rate. Therefore, the organization can invest in different kinds of keywords by following their current strategy or for the purpose to fulfill the urgent needs. User device category is crucial in the first two phases
(acquisition and activation), yet it seems not very important when it comes to retention and revenue. As for the language difference, it should not be the top priority to consider when buying a keyword.
The significances and the impacts (“+” as positive, “–” as negative) of the independent variables for each stage in the AARRR model are presented in Table 25.
Table 25: Summary of regression analysis in AARRR model
Stage Acquisition Activation Retention Revenue
Variables Click BounceRate %NewSessions Conversion
foundation + *** –** –***
opensource –** –*** + **
tech –·
gov
event_cs + **
event_camp + ** –·
action_donation
action_newsletter + * –*
device_mobile –* + ***
device_tablet –*** + ***
language_ch + **
language_mix
Note. Significant codes: *** p<0.001, ** p<0.01, * p<0.05, · p<0.1.
We summarize the implications in the tables below. Only the significant attributes (p-value < 0.1) are listed, with the mark “+” as effective and worth investing and the mark “–” as the indicator of poor performance and need to be adjusted.
Table 26: Summary of implications and suggestions
Acquisition Activation Retention Revenue
Keyword worth keep investing, except for the government-related ones, whose relationships with the measurements are as not satisfying.
Event Type
event_cs (+)
event_camp (–) event_camp (+)
Reinforce the computer science-related event promotion in AdWords.
Forsake the clickbait-like ones in student camp category, for they perform poorly both in acquisition and activation. (Although they work well in retention, the reason needs further investigation for determination.)
Action Type
newsletter (+) newsletter (+)
Pursuit awareness and exposure by newsletter keywords. Set a new goal of “newsletter subscription” in Google Analytics are also suggested, so the relationships between retention and subscription can be examined.
donation has no strong impact throughout the stages. Keyword advertising may not be the best way for conversion.
Table 27: Summary of implications and suggestions (continued)
Acquisition Activation Retention Revenue
Device
mobile (–) tablet (–)
mobile (–) tablet (–)
Mobile and tablet users seem to click less and leave the sites quickly on average than desktop users. In addition, there is room for improvement for all the device categories in OCF comparing to the benchmark.
Language
Chinese (–)
Language is the last priority to consider when buying keywords.
English keywords have lower bounce rate than Chinese ones, might because most English keywords are about specific technologies, and better cater the target audience.
Chapter 5 Conclusions
5.1 Research Purpose and Contribution
Many small NPOs share the common struggle of the shortage of funds when it comes to marketing. Thus, they need to strive to seize every resource and opportunity.
Google Ad Grants program is one of the opportunities that widely used by NPOs globally, which provides free in-kind keyword advertising service. Nevertheless, even given the considerable budgets, how to select the right keyword wisely remains a great challenge faced by advertisers.
In this thesis, our research purpose is to reveal the possibilities for small NPOs to better utilize the limited resource and optimize their online marketing performance. In order to make use of the opportunity given from the Google Ad Grants program, we accordingly focus on the keyword advertising. Our goal is to help NPOs identify the best-suited and most effective keywords. Specifically, we wish to find out the particular attributes of the keywords that can affect advertising performance in different phases of customer lifecycle and conversion behavior designated by the AARRR model.
The contribution of our thesis is the pioneering research in the field of Taiwan-based NPO focusing on the interplay between keyword attributes and advertising performance. The study provides guidelines for future researches and practitioners regarding the impact of keyword attributes in different phases in customer lifecycle and conversion behavior. From an academic perspective, our attempts to uncover the
improvement opportunities for Taiwanese NPO and to explain the relationships between keyword attributes and advertising effectiveness is innovative as well as pioneering. The reasoning behind attribute category and variable designation, the statistical model selection, and the approaches to analyze and interpret are all adoptable for future studies. From an advertiser’s point of view, our findings can help them to re-examine their existing keywords, and enable them to select and invest in keywords based on a mathematically-proved methodology, rather than mere intuitions.
5.2 Findings and Suggestions
In our empirical analysis, we designate 12 keyword attributes in five categories: 1) Keyword Essence: foundation, open source, technology, government; 2) Event Type:
computer science event, student camp event; 3) Action Type: call for donation, call for newsletter subscription; 4) Device Category: mobile, tablet (desktop as reference level);
5) Language: Chinese, mixed language (English as reference level). The relationships between the attributes and the phases of customer lifecycle and conversion behavior – acquisition, activation, retention, referral – are being observed and explained.
In the category of keyword essence, the overall outcome is promising throughout all the stages, indicating that the current keyword attributes are worth investing as well as keeping exploring, only except for the government-related ones, however, whose relationships with the measurements are as not satisfying.
In terms of event type, the promotion of the computer science-related event in keyword advertising should be reinforced, based on the positive results in visitor attracting. On the other hand, we recommend the organization to forsake or adjust the clickbait-like ones in the student camp category, for they perform poorly both in acquisition and activation stage.
As for call for action keywords, newsletter has strong tendency to garner clicks, making them good choices to build awareness. They also work well in retention; we suggest setting a goal of “newsletter subscription” in Google Analytics in order to track and understand the relationships between retention and subscription. Keyword
advertising may not be the best way to drive conversions, for donation attribute has no significant impact throughout the stages.
Device-wise, mobile and tablet users seem to click less and leave the sites quickly on average than desktop users. Yet there is no obvious difference in device when it comes to retention and revenue phase. Additionally, in the perspective of user experience, there is room for improvement for all the device categories in OCF comparing to the benchmark.
Language seems to be the last priority to consider when buying keywords, which hardly has significant influences. The only exception is the activation stage, where English outperforms others; one possible explanation is that most English keywords are about specific technologies, and thus better cater the target audience.
To summarize, the research results indicate that some attributes have strong impacts throughout most of the stages, which makes them good options for investment.
Meanwhile, some attributes are only influential in particular phases. Therefore, the organization should prioritize their keyword selection by following their current strategy or to fulfill the urgent needs.
5.3 Limitation and Future Work
Some different adaptations, tests, comparisons, and analysis have been left for the future due to the time limit and resource constraints. First, our regression models focus on the relationship between each independent variable (keyword attribute) and the dependent variable (performance measurement), which is the main effect. Among the keyword attributes, we are aware that the independent variables might interact with each other in a more complex reality. Therefore, the interaction effects can be taken into account when conducting the further analyses.
Secondly, user behavior, such as bounce rate or re-visit, can be affected by the website content, interface design, loading time, or other experience. In our discussion, we solely focus on the keywords that the users click on, and we believe that reflects user’s intention at the moment, but we are also aware of the limitation; hence, we can further conclude the analysis of website content, such as user research, eye tracking heatmaps, etc.
In terms of research resource, with access to more data, we can try to discover measurements that interpret the relation between keyword advertising and the referral
behavior or the referral outcome. In addition, in our case study, there is only one goal,
“donation process completion”, set in Google Analytics. With more goals designed, such as “newsletter subscription” or “media kit download”, we can have a broader view on the overall marketing performance.
Lastly, we have identified the potential influential keyword attributes in the
empirical analysis. Based on our suggestions, we can conduct follow-up experiments by changing keyword strategy and carrying out continuous A/B tests, and thus verify the research findings.
References
1. Anheier, H. K. (2005). Nonprofit Organizations: Theory, Management, Policy (1st ed.). Oxon: Routledge.
2. Belch, G. E., & Belch, M. A. (2004). Advertising and Promotion: An Integrated Marketing Communications Perspective (6th ed.). Boston: Irwin.
3. Blake, T., Nosko, C., & Tadelis, S. (2015). Consumer Heterogeneity and Paid Search Effectiveness: A Large‐scale Field Experiment. Econometrica, 83(1), 155–
174.
4. Bruce, I. (1995). Do not-for-profits value their customers and their needs ? International Marketing Review, 12(4), 77–84.
5. Chaffey, D. (2006). Internet marketing: strategy, implementation and practice.
Harlow: Financial Times Prentice Hall.
6. Chaffey, D., & Smith, P. R. (2005). Marketing Excellence: at the Heart of E-Business (2nd ed.). Oxford: Butterworth Heinemann.
7. Chambers, J. M. (1992). Linear Models. In J. M. Chambers & T. J. Hastie (Eds.), Statistical Models in S. Wadsworth & Brooks/Cole.
8. Chris, A. (2015). 5 Reasons Why PPC is important for Small Business Success.
Retrieved December 9, 2017, from https://www.digitalmarketingpro.net/5-reasons-why-ppc-is-important-for-small-business-success-550/
9. CMI, & Blackbaud. (2015). Nonprofit Content Marketing - 2015 Benchmarks, Budgets and Trends - North America. Retrieved December 8, 2017, from
https://www.slideshare.net/CMI/nonprofit-content-marketing-2015-benchmarks- budgets-and-trends-north-america-by-cmi-blackbaud-and-sponsored-by-fusion-spark
10. Edelmen, D. C. (2010). Branding in the Digital Age: You’re Spending Your Money in All the Wrong Places. Harvard Business Review, (December).
11. Eisenhardt, K. M. (1989). Building Theories from Case Study Research. Academy of Management Review, 14(4), 532–550.
12. Eley, B., & Tilley, S. (2009). Online Marketing Inside Out (1st ed.). SitePoint.
13. Ellis, S., & Brown, M. (2017). Hacking Growth: How Today’s Fastest-Growing Companies Drive Breakout Success. Crown Business.
14. eMarketer. (2017). Google and Facebook Tighten Grip on US Digital Ad Market.
Retrieved December 9, 2017, from https://www.emarketer.com/Article/Google-Facebook-Tighten-Grip-on-US-Digital-Ad-Market/1016494
15. Ghose, A., & Yang, S. (2009). An Empirical Analysis of Search Engine Advertising:
Sponsored Search in Electronic Markets. Management Science, 55(10), 1605–1622.
16. Google Ad Grants. (2017). Success Stories. Retrieved December 9, 2017, from https://www.google.com/grants/success-stories/#?modal_active=none
17. Guz, D. De. (2016). What Is A Growth Hacker and Why It’s the Fastest Growing Mmarketing Role at Startups. Retrieved December 3, 2017, from
https://blog.rebrandly.com/growth_hacker/
18. Hart, T. R. (2002). ePhilanthropy : Using the Internet to build support. Journal of Nonprofit & Voluntary Sector Marketing, 7(4), 353–360.
19. Hastie, T. J., & Pregibon, D. (1992). Generalized Linear Models. In J. M. Chambers
& T. J. Hastie (Eds.), Statistical Models in S. Wadsworth & Brooks/Cole.
20. Hatch, J. (2015). The Ad Rank Formula Revealed. Retrieved from
https://www.disruptiveadvertising.com/adwords/the-ad-rank-formula-revealed/
21. Holiday, R. (2013). What Is Growth Hacking? A Definition and a Call to Action.
Retrieved December 1, 2017, from http://www.huffingtonpost.com/ryan-holiday/what-is-growth-hacking-a-_b_3863522.html
22. Holiday, R. (2014). Growth Hacker Marketing: A Primer on the Future of PR, Marketing, and Advertising. Portfolio.
23. Ingenhoff, D., & Koelling, A. M. (2009). The potential of Web sites as a relationship building tool for charitable fundraising NPOs. Public Relations Review, 35(1), 66–
73. https://doi.org/10.1016/j.pubrev.2008.09.023
24. Jerath, K., Ma, L., & Park, Y. H. (2014). Consumer Click Behavior at a Search Engine: The Role of Keyword Popularity. Journal of Marketing Research, 5(14), 480–486.
25. Kotler, P. (1979). Strategies for introducing marketing into non-profit organizations.
Journal of Marketing, 43(1), 37–44.
26. Li, H. (Alice), Kannan, P. K., Viswanathan, S., & Pani, A. (2016). Attribution Strategies and Return on Keyword Investment in Paid Search Advertising.
Marketing Science, 35(6), 831–848. https://doi.org/10.1287/mksc.2016.0987 27. Liquidreach. (2014). How Much Should A Non-Profit Spend On Marketing?
Retrieved December 8, 2017, from http://www.royalladv.com/blog/non-profit-marketing-budget/
28. McClure, D. (2007). Product Marketing for Pirates: AARRR! (aka Startup Metrics for Internet Marketing & Product Management). Retrieved September 15, 2017, from http://500hats.typepad.com/500blogs/2007/06/internet-market.html
29. McPherson, R. C. (2007). Digital Giving: How Technology is Changing Charity.
iUniverse, Inc.
30. Merriam-Webster. (2017). Nonprofit. Retrieved September 10, 2017, from
https://www.merriam-webster.com/dictionary/nonprofit
31. Molla, R. (2017). Google and Facebook are Driving Nearly All Growth in the Global Ad Market.
32. Netzer, Y. (2011). Keyword Optimization in Search-Based Advertising Markets. Tel Aviv University.
33. Opreana, A., & Vinerean, S. (2015). A New Development in Online Marketing : Introducing Digital Inbound Marketing, 3(1), 29–34.
34. Pan, Y., & Jackson, R. T. (2008). Ethnic Difference in the Relationship between Acute Inflammation and Serum Ferritin in US Adult Males. Epidemiology and Infection, 136(3), 421–431.
35. Patel, N. (2015). Growth Hacking Made Simple: A Step-by-Step Guide. Retrieved December 1, 2017, from https://neilpatel.com/what-is-growth-hacking/
36. Peters, R. (2014). Growth Hacking Techniques, Disruptive Technology - How 40 Companies Made It BIG. World Ideas.
37. Pinho, J. C., & Macedo, I. M. (2016). The Benefits and Barriers Associated with the Use of the Internet Within the Non-Profit Sector. Journal of Nonprofit & Public Sector Marketing, 16(1–2), 171–193.
38. PwC, & Interactive Advertising Bureau. (2017). IAB Internet Advertising Revenue Report.
39. Rathi, D., Given, L. M., Forcier, E., Rathi, D., Given, L. M., & Forcier, E. (2016).
Knowledge needs in the non-profit sector : an evidence-based model of organizational practices. Journal of Knowledge Management, 20(1), 23–48.
https://doi.org/10.1108/JKM-12-2014-0512
40. Rogerson, P. A. (2001). Statistical Methods for Geography. London: Sage
Publications Ltd.
41. Rutz, O. J., & Bucklin, R. E. (2011). From Generic to Branded: A Model of
Spillover in Paid Search Advertising. Journal of Marketing Research, 48(1), 87–102.
42. Saxton, G. D., Guo, C., Brown, W. A., Taylor, P., Saxton, G. D., & Brown, W. A.
(2007). Based Technologies NEW DIMENSIONS OF The Application and Promise of. Public Performance & Management Review, 31(2), 144–173.
https://doi.org/10.2753/PMR1530-9576310201
43. Saxton, J. (2001). The growth of the Internet , digital television and mobile
telephony and the implications for not-for-profit marketing. Nonprofit and Voluntary Sector Quarterly, 6(4), 347–363.
44. Socailsquare. (2014). What is Open Source explained in LEGO. Retrieved September 1, 2017, from https://www.youtube.com/watch?v=a8fHgx9mE5U 45. Vunk, R. (2017). Growth Hacking in Estonian Start-ups. Estonian Business School.
46. Yang, X., Deng, T., Guo, Z., & Ding, Z. (2017). Advertising Keyword Recommendation based on Supervised Link Prediction in Multi-Relational Network. WWW (Companion Volume), 863–864. Retrieved from http://dblp.uni-trier.de/db/conf/www/www2017c.html#YangDGD17
47. Yin, R. K. (1994). Study Design and Methods. Applied Social Research Method Series (2nd ed., Vol. 5). SAGE publications.
48. Yorke, D. A. (1984). Marketing and Non-Profit-Making Organisations. European Journal of Marketing, 18(2), 17–22.