Skip to main content
SpringerLink
Log in

Price promotions and “freemium” app monetization

  • Published:
Quantitative Marketing and Economics Aims and scope Submit manuscript

Abstract

The “freemium” model for digital goods involves selling a base version of the product for free, and making premium product features available to users only on payment. The success of the model is predicated on the ability to profitably convert free users to paying ones. Price promotions (or “sales”) are often used in freemium to induce this conversion. However, the causal effect of exposing consumers to such inter-temporal price variation is unclear. While sales can generate beneficial short-run conversion, they may be harmful in the long-run if consumers inter-temporally substitute purchases to periods with low prices, or use them as signals of low product quality. These long-run concerns may be accentuated in freemium apps, where the base version is sold for free, so that sales form extreme price cuts on the overall product combination. We work with the seller of a free-to-play video game to randomize entering cohorts of users into treatment and control conditions in which promotions for in-game purchases are turned on or off. We observe complete user behavior for half a year, including purchases and consumption of in-game goods, which, in contrast to much of the extant literature, enables us to assess possible substitution over time in consumption directly. We find that conversion and revenue improve in the treatment group; and detect no evidence of harmful inter-temporal substitution or negative inferences about quality from exposure to price variation, suggesting that promotions are profitable. We conjecture that the zero price of the base product that makes its consumption virtually costless, combined with the complementarity between the base product and premium features can help explain this. To the extent that this holds across freemium contexts, the positive effects of promotions documented here may hold more generally.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9

Similar content being viewed by others

Notes

  1. The difficulty of converting free users to paying ones could arise from a fixed cost to the user of setting up payment (e.g., the customer has to enter his credit card details and get verified) or by the special significance of a “zero price” in consumer’s minds (Heyman & Ariely, 2004; Shampanier et al., 2007; Ascarza et al., 2012).

  2. In the context of gaming apps, we can think of the base version of the game as the base good and the premium features as the add-on goods. Spending in-game currency allows the user to acquire add-on goods.

  3. In-game purchases in free-to-play games accounted for 82% of worldwide digital games revenue in 2017 (Gough, 2018). The market for mobile in-game consumer spending alone makes up close to 80% of $101 billion in-app revenue in 2018 (App Annie, 2018).

  4. See https://en.wikipedia.org/wiki/Puzzle_video_game for descriptions of this genre.

  5. A small amount of cash can be earned in rare situations within the game, such as when solving a specially designed puzzle. However, such situations are highly infrequent and the cash earned this way is minuscule. Such currency configurations are typical for the app game genre (e.g., see https://en.wikipedia.org/wiki/Virtual_economy#Types_of_virtual_currencies, accessed Dec 22, 2020).

  6. It is also representative of the “app economy” where advertising only drives a small share of revenue (Ghose and Han, 2014).

  7. A treated user is exposed to the treatment for 180 days. A user who drops off is exposed to the treatment for ≤ 180 days. Thus, the intensity of treatment is lesser for the attrited group. We measure an intent-to-treat effect (ITT), comparing all users who were randomized into treated versus control groups (including those that drop off). The noise in the ITT is higher the higher the number of users who drop off.

  8. The following filters are applied for data cleaning. Data are observed at the device-level. About 50% of devices are connected to one or more Facebook accounts. We disregard all devices that are connected to more than one Facebook account as this indicates the game was played by more than one user on the respective device. We further disregard devices that are connected to Facebook accounts that previously played the game on other platforms, and we disregard devices with Facebook accounts that were connected to multiple devices during the experiment. The remaining sample contains devices for which we can reasonably assume that only one user played the game on them. This data was further cleaned for duplicate and obviously broken log entries and for devices that experienced technical difficulties or where the user “hacked” the app during the experimental period.

  9. Thanks to a referee for suggesting this.

  10. Figures 12 and 13 in the Appendix document that there is no difference in the propensity to log in between the treatment and control groups.

  11. Note that the regression is not exact and ρi can sometimes be greater than one because there is no way for us to track reliably whether the cash spent during the promotion is derived fully from the cash bought during promotions. For instance, all users start with an initial endowment of all currencies (including cash), and some of that endowment could be redeemed during a promotional period over and above the cash bought during those periods.

  12. We report these for cash rather than energy or coins because its interpretation is cleaner, as discussed above and also later in this section.

  13. The reverse could happen in a perverse situation where those that value highly playing the base game value its premium features less. Then, valuations would be negatively, rather than positively, covarying across users.

  14. We also see in Fig. 9 that advertising revenue is lower when price promotions begin earlier. Given the patterns in Fig. 8, this is not driven by a mechanical linkage of advertising exposure and advertising revenue to time spent in the app. The fact that early onset of promotions reduces advertising revenue without changing time spent in the app significantly, suggests that promotions and advertising are substitutes from the perspective of users. This can occur for instance, because both provide ways to obtain in-game currencies, or because both compete for scarce user attention. There is some evidence in the literature that promotions can serve as advertising (e.g., Sahni et al., 2016 and the literature cited there), so it is possible that promotions tend to deplete a “mental account” associated with attention to advertising.

References

  • Alter, A. (2017). Irresistible: The rise of addictive technology and the business of keeping us hooked. United Kingdom: Penguin Books.

    Google Scholar 

  • Anderson, E.T., & Simester, D.I. (2004). Long-run effects of promotion depth on new versus established customers: three field studies. Marketing Science, 23, 4–20.

    Article  Google Scholar 

  • Anderson, E.T., & Simester, D.I. (2010). Price stickiness and customer antagonism. The Quarterly Journal of Economics, 125, 729–765.

    Article  Google Scholar 

  • App Annie. (2018). The state of mobile 2019, App Annie Blog. https://www.appannie.com/en/go/state-of-mobile-2019/, (Accessed May 18 2020).

  • Appel, G., Libai, B., Muller, E., & Shachar, R. (2020). On the monetization of mobile apps. International Journal of Research in Marketing, 37, 93–107.

    Article  Google Scholar 

  • Ascarza, E., Lambrecht, A., & Vilcassim, N. (2012). When talk is ’free’: the effect of tariff structure on usage under two- and three-part tariffs. Journal of Marketing Research, 49, 882–899.

    Article  Google Scholar 

  • Bell, D.R., & Lattin, J. (1998). Shopping behavior and consumer preference for retail price format: why “large basket” shoppers prefer EDLP. Marketing Science, 17, 66–88.

    Article  Google Scholar 

  • Carter, B. (2019). Freemium conversion issues? Why you need to address the penny gap, GoSquared Blog. https://www.gosquared.com/blog/freemium-conversion-issues, (Accessed May 18 2020).

  • Cho, S., & Rust, J. (2010). The flat rental puzzle. The Review of Economic Studies, 77, 560–594.

    Article  Google Scholar 

  • Dubé, J.-P., Luo, X., & Fang, Z. (2017). Self-signaling and prosocial behavior: a cause marketing experiment. Marketing Science, 36, 161–186.

    Article  Google Scholar 

  • Dube, J.-P., & Misra, S. (2017). Scalable price targeting. Working paper, Booth School of Business.

  • Elberg, A., Gardete, P., Macera, R., & Noton, C. (2019). Dynamic effects of price promotions: field evidence, consumer search, and Supply-Side implications. Quantitative Marketing and Economics, 17, 1–58.

    Article  Google Scholar 

  • Ellickson, P.B., & Misra, S. (2008). Supermarket pricing strategies. Marketing Science, 27, 811–828.

    Article  Google Scholar 

  • Ellickson, P.B., Misra, S., & Nair, H.S. (2012). Repositioning dynamics and pricing strategy. Journal of Marketing Research, 49, 750–772.

    Article  Google Scholar 

  • Erdem, T., Imai, S., & Keane, M.P. (2003). Brand and quantity choice dynamics under price uncertainty. Quantitative Marketing and Economics, 1, 5–64.

    Article  Google Scholar 

  • Erdem, T., Keane, M.P., & Sun, B. (2008). A dynamic model of brand choice when price and advertising signal product quality. Marketing Science, 27, 1111–1125.

    Article  Google Scholar 

  • Eyal, N. (2014). Hooked: how to build habit-forming products. United Kingdom: Penguin Books.

    Google Scholar 

  • Fyber. (2020). 2019 State of in-app advertising and monetization. Fyber. https://blog.fyber.com/2019-%20state-of-in-app-advertising-and-monetization/, (Accessed March 18 2021).

  • Gerstner, E. (1985). Do higher prices signal higher quality? Journal of Marketing Research, 22, 209–215.

    Article  Google Scholar 

  • Ghose, A., & Han, S.P. (2014). Estimating demand for mobile applications in the new economy. Management Science, 60, 1470–1488.

    Article  Google Scholar 

  • Gough, C. (2018). Distribution of digital games market revenue worldwide in 2017, by monetization model. Statista - The Statistics Portal. https://www.statista.com/statistics/821451/distribution-digital-games-market-revenue-monetization-model/, (Accessed May 18 2020).

  • Hendel, I., & Nevo, A. (2003). The post-promotion dip puzzle: what do the data have to say? Quantitative Marketing and Economics, 1, 409–424.

    Article  Google Scholar 

  • Heyman, J., & Ariely, D. (2004). Effort for payment: a tale of two markets. Psychological Science, 15, 787–793.

    Article  Google Scholar 

  • Ho, T.-H., Tang, C.S., & Bell, D.R. (1998). Rational shopping behavior and the option value of variable pricing. Management Science, 44, 145–160.

    Article  Google Scholar 

  • Hoch, S., Dreze, X., & Purk, M. (1994). EDLP, Hi-Lo, and margin arithmetic. Journal of Marketing, 58, 16–27.

    Article  Google Scholar 

  • Jo, W., Sunder, S., Choi, J., & Trivedi, M. (2020). Protecting consumers from themselves: assessing consequences of usage restriction laws on online game usage and spending. Marketing Science, 39, 117–133.

    Article  Google Scholar 

  • Levitt, S.D., List, J.A., Neckermann, S., & Nelson, D. (2016). Quantity discounts on a virtual good: the results of a massive pricing experiment at king digital entertainment. Proceedings of the National Academy of Sciences, 113, 7323–7328.

    Article  Google Scholar 

  • Mela, C.F., Gupta, S., & Lehmann, D.R. (1997). The long-term impact of promotion and advertising on consumer brand choice. Journal of Marketing Research, 248–261.

  • Milgrom, P., & Roberts, J. (1986). Price and advertising signals of product quality. Journal of Political Economy, 94, 796–821.

    Article  Google Scholar 

  • Nair, H.S., Misra, S., Hornbuckle, W.J., Mishra, R., & Acharya, A. (2017). Big data and marketing analytics in gaming: combining empirical models and field experimentation. Marketing Science, 36, 699–725.

    Article  Google Scholar 

  • Neslin, S.A., & van Heerde, H.J. (2009). Promotion dynamics. Foundations and Trends®; in Marketing, 3, 177–268.

    Article  Google Scholar 

  • Nevskaya, Y., & Albuquerque, P. (2019). How should firms manage excessive product use? A continuous-time demand model to test reward schedules, notifications, and time limits. Journal of Marketing Research, 56, 379–400.

    Article  Google Scholar 

  • Nijs, V.R., Dekimpe, M.G., Steenkamps, J.-B.E., & Hanssens, D.M. (2001). The category-demand effects of price promotions. Marketing Science, 20, 1–22.

    Article  Google Scholar 

  • Rao, A.R., & Monroe, K.B. (1989). The effect of price, brand name, and store name on buyers’ perceptions of product quality: an integrative review. Journal of Marketing Research, 26, 351–357.

    Google Scholar 

  • Rubin, D.B. (1978). Bayesian inference for causal effects: the role of randomization. The Annals of Statistics, 6, 34–58.

    Article  Google Scholar 

  • Sahni, N.S., Zou, D., & Chintagunta, P.K. (2016). Do targeted discount offers serve as advertising? Evidence from 70 field experiments. Management Science, 63, 2688–2705.

    Article  Google Scholar 

  • Sensortower. (2020). Q2 2019 store intelligence data digest. Sensortower. https://sensortower.com/reports, (Accessed March 18 2021).

  • Shampanier, K., Mazar, N., & Ariely, D. (2007). Zero as a special price: the true value of free products. Marketing Science, 26, 742–757.

    Article  Google Scholar 

  • Shmilovici, U. (2011). The complete guide to freemium business models. Techcrunch. https://techcrunch.com/2011/09/04/complete-guide-freemium/, (Accessed May 18 2020).

  • Tomic, N.Z. (2019). Economic model of micro transactions in video games. Journal of Economic Science Research, 1.

Download references

Author information

Authors and Affiliations

  1. Fuqua School of Business, Duke University, 100 Fuqua Drive, Durham, NC, 27708, USA

    Julian Runge

  2. Institute for Marketing, RWTH Aachen University, Kackertstr. 7, 52072, Aachen, Germany

    Julian Runge

  3. Stanford Graduate School of Business, Stanford University, 655 Knight Way, Stanford, CA, 94305, USA

    Jonathan Levav & Harikesh S. Nair

Authors
  1. Julian Runge
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Jonathan Levav
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Harikesh S. Nair
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Julian Runge.

Ethics declarations

Conflict of Interests

The first author worked in part-time employment as a research scientist at the industry collaborator while data for the study were collected.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

We thank the video game company on whose mobile video game app we implemented the experiment reported in this paper. The company wishes to remain anonymous. No confidential information, user details or trade secrets are revealed in this paper. All opinions in this paper represent the views of the authors and not of the company or the respective universities.

Appendix

Appendix

1.1 In-game currency and sales in Candy Crush Saga

Fig. 10
figure 10

Examples of In-Game Currency and Sales in the Popular Video Game Candy Crush Saga. The left panel shows a screen shot of “gold bars”, a type of in-game currency available for purchase in the game Candy Crush Saga. The user can buy the bars in various bundles. The right panel shows a sale run in the game for a bundle of in game currencies and other in-game features

Full size image

1.2 Days in app is balanced between treatment and control groups

Fig. 11
figure 11

Q-Q Plot of Days to Complete Level One in Treated and Control Groups. Users in the treatment group have to complete level one before they are exposed to the promotional schedule. Since users are not exposed to any promotions before they reach level one, the set of users who reach level one in the treatment and control group remain balanced. Suppose a user i in the treated group reaches level one in τi days, so is exposed to the treatment for di = 180 − τi days. The figure presents a quantile-quantile plot of the distribution of di in the treatment (y-axis) and control (x-axis) groups to check balance. The distribution of di in the treatment and control groups is seen to be balanced

Full size image

1.3 Balance tests: Treatment vs. control groups

Table 7 Pre-treatment Tests of Balance for Treatment and Control Groups
Full size table

1.4 Balance tests: Three treatment sub-groups

Table 8 Pre-treatment Tests of Balance for Treatment Sub-groups
Full size table
Fig. 12
figure 12

Number of Users Logging into the App by Day Split by Group

Full size image
Fig. 13
figure 13

Q-Q Plot of Number of Days a User is Active in the App Split by Group. Days active is defined as the number of days (out of 180) a user logged into the game. The distribution of login propensity is seen to be the same across the treatment and control groups

Full size image
Fig. 14
figure 14

Per-day Purchases for Treated and Control Groups. Purchases per day with 95% confidence band over calendar days. Grey diamonds on the date axis indicate days with a promotional cycle

Full size image
Fig. 15
figure 15

Per-day Revenues for Treated and Control Groups. Revenue per day with 95% confidence band over calendar days. Grey diamonds on the date axis indicate days with a promotional cycle

Full size image

1.5 Information about sales is widely discussed on online gaming forums

Fig. 16
figure 16

Screen Shots from a Gaming Forum where Information about Sales on the Video Game Is Discussed Prior to the Experiment

Full size image

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Runge, J., Levav, J. & Nair, H.S. Price promotions and “freemium” app monetization. Quant Mark Econ 20, 101–139 (2022). https://doi.org/10.1007/s11129-022-09248-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11129-022-09248-3

Keywords

JEL Classification

Search

Navigation