Innovation management in crisis: patent analytics as a response to the COVID-19 pandemic

Crises like the COVID-19 pandemic affect firms’ innovation management and decision making. On the downside, crises lead to detriments like budget constraints, to which firms often respond by reducing their innovation activities. On the upside, crises are opportunities, where some firms exploiting changing market requirements and necessities excel. No matter in which direction, decision makers must react quickly but often rely on ad-hoc decisions or even gut feeling when drafting their crisis response strategies. Through a series of distinct cases, we demonstrate that innovation management may fill this void through patent analytics. Drawing on biochemical expertise, we particularly describe the functions and effects of COVID-19. To counter downside detriments, firms may circumvent budget constraints by discerning patents that can be (1) monetized, for example via sales or licensing deals, or (2) abandoned to achieve cost-savings, allowing firms to maintain their innovation activities. To realize upside opportunities, firms and governments may use patent analytics to detect key biotechnology firms that are likely to successfully develop treatments and vaccinations against pandemics like COVID-19. Promulgated U.S. interest in relocating foreign firms to the United States is not without technological and commercial reasoning. Herein, the insights of this study contribute to a better understanding of the use of patent information, such as smart patent indicators, harmonized patent data, novel annuity fee measures, and hand-collected datasets of COVID-19 and related antibodies’ patents to the management of innovation in times of crisis


Introduction
T he recent outbreak of the novel virus SARS-CoV-2 has developed into a worldwide pandemic from late 2019 onwards (e.g., Rothan and Byrareddy, 2020). Firms all over the world suffer from economic detriments due to lockdown-induced demand reductions, resulting in budget constraints, uncertain innovation prospects, and unpredictable future developments (Paunov, 2012;Tietze et al., 2020). Whilst most prior crises were limited to specific industrial sectors, countries, or firms, the COVID-19 pandemic is truly global, affecting everyone.
On the downside, prior research investigating innovation management in crises found crises to intensify 'uncertainty, complexity, ambiguity and 2 R&D Management 2020 unpredictability' (Davis et al., 2009;Martin-Rios and Pasamar, 2018). Corporate decision makers have to react quickly, consider future effects, and simultaneously generate options to overcome the crisis here and now (Teece et al., 2016). These reactions are typically ad hoc and based on gut feelings instead of sufficient information (e.g., Müller, 1985;Cooper and Edgett, 2010;Bessant et al., 2015). For example, firms often cut their innovation investments despite negative long-term consequences, as visible in reduced patent filings, forfeiting the firms' future (Archibugi et al., 2013b;Hingley and Park, 2017).
On the upside, prior research since Schumpeter's (1934) seminal work simultaneously demonstrated that crises-induced financial constraints foster innovation (Hoegl et al., 2008). For example, Archibugi et al. (2013a) found that whilst crises lead to concentrations of innovative activities for growing and innovative firms, expanding explorative strategies enable firms to better cope with and overcome crises. In addition to those firms that suffer from the crises, some winners emerging stronger also exist (Archibugi et al., 2013b). For example, firms like AirBnB and Uber emerged like a phoenix arising from the ashes after the recent financial crisis, with innovative offerings matching crises' Zeitgeist and market requirements (e.g., Oskam and Boswijk, 2016;Zervas et al., 2017). However, success stories become success stories ex-post only, not when they begin.
In fact, it is challenging and often a mere chance to successfully bet on future winners without sufficient data to back decisions, and even more difficult though in crisis situations (e.g., Sah and Stiglitz, 1986;Christensen and Knudsen, 2010;Denrell and Fang, 2010). It is surprising that only little scholarly attention has been paid to investigating the possibilities that publicly available information offers for organizational decision making, not only in the global COVID-19 pandemic, but in crises in general. In this paper, we address this gap by studying how innovation management decisions in times of crisis like the COVID-19 pandemic can be improved through publicly available data.
To address this research question, we turn to patent analytics (Campbell, 1983;Ashton and Sen, 1988;Ernst, 2003). We show how patent analytics may shape innovation management during crises (e.g., Campbell, 1983;Archibugi et al., 2013a;Guderian, 2019). By drawing on longitudinal and crosssectional patent data, we present four distinct case studies. The first two cases address the downside, i.e., how locked-down firms can reduce expenses, free budget to continue innovation efforts, and may generate excess revenue based on data-driven decisions, for example by identifying technology licensing and abandonment candidates. The latter two cases address the upside, i.e., how success stories are identifiable ahead of time using data on their technologies and innovation activities, for example, to detect the most promising candidates to develop COVID-19 treatments and vaccinations.
The insights of this study yield several contributions to the literature on innovation management in times of crises and patent analytics in innovation management (e.g., Döner, 2017;Jung et al., 2018;Tietze et al., 2020;Antonioli and Montresor, 2019). We show that decision makers can use patent information and analytics to find data-driven innovation management solutions to overcome crisis situations, as in the COVID-19 pandemic. This includes the firm-internal view on own technologies and innovation activities as well as the firm-external view on incumbent and novel competition and technology evolvement. Using cross-disciplinary insights from natural sciences and strategic patent management, we identify COVID-19 and related immunology-based' patent families, detect key biotechnology firms already possessing experience in developing cures, treatment, and vaccination to related coronavirus diseases using smart patent indicators and harmonized patent data. Moreover, we highlight why a promulgated U.S. interest in relocating firms like CureVac might not be without technological and commercial reasoning. Further, we demonstrate how firms like Adidas can optimize their budget constraints by detecting patent families that can be monetized, e.g., via sales or licensing deals, as well as potentially saving costs by focusing on impactful patent families and novel annuity fee measures. This provides firms with options to maintain innovative activities in crisis situations. Taken together, these case studies show how patent analytics support firms' managements in mitigating crises like the COVID-19 pandemic and thus enabling them to emerge stronger eventually as a success story. Herein, we provide concrete recommendations for management and depict arrays for future research.

Innovation management in crises
Crises influence innovation management (Döner, 2017;Teplykh, 2018;Antonioli and Montresor, in press). On the downside, prior research found crises to intensify 'uncertainty, complexity, ambiguity and unpredictability' (Davis et al., 2009;Martin-Rios and Pasamar, 2018;Teplykh, 2018). Compared to favorable contexts, in which innovation thrives, innovation is hindered in more unfavorable scenarios, creating innovation barriers, as in crises (Ferreira and Teixeira, 2016;Teplykh, 2018). For example in the financial crises of the late 2000', firms changed their innovation behavior (Cruz-Castro et al., 2018). Disoska et al. (2020) found an indication that crises have a negative influence on firms' willingness to innovate. Some firms only maintain their innovation activities in exchange for public support, which implies a de-facto reduction in proprietary innovation efforts (Antonioli and Montresor, in press). Decreases in firms' revenues directly affect research and development, hence innovation outputs (Döner, 2017).
On the upside, crises do not only entail adverse effects, but offer opportunities stimulating innovation (Carmeli and Schaubroeck, 2008;Ulmer et al., 2011;Cefis and Marsili, 2019;Heyden et al., 2020). When firms manage to meet the shifting market requirement and crises-induced necessities, they can emerge even stronger from these situations (Archibugi et al., 2013a;Mayr et al., 2017). For example, Nemlioglu and Mallick (2020) interpret innovation as a path out of crisis-induced valuation uncertainty, where financially less-constrained, innovative firms yield higher values and experience less uncertainty. In the same vein, Heyden et al. (2020) identified that corporate management benefits from enabling radical change initiatives instead of cutting costs. Innovative activity may support firms in shielding themselves from crises' effects, while other supposed resilience fostering actions, such as marginal production cost reductions, fail (Gupta, 2019). Cefis et al. (2020) found that innovations grant a survival premium. Regardless of firms' financial conditions, innovative firms have higher probabilities than non-innovative firms to survive crises (Cefis et al., 2020).

Innovation management and patent analytics
In addition to the link between crises and innovation management, prior research has also established a relation between innovation management and patent analytics (Candelin-Palmqvist et al., 2012;Holgersson, 2013). To profit from innovation, firms often rely on intellectual property rights like patents to appropriate returns (Greenhalgh et al., 2001;Somaya, 2012;Di Minin and Faems, 2013;Bican et al., 2017). In industries like pharmaceuticals, this is particularly prevalent (Hemphill and Sampat, 2012;Conley et al., 2013a;DiMasi et al., 2016).
Patents are publicly available and serve as objective data sources on firms' innovation activities (Ashton and Sen, 1988;Buehler et al., 2017;Guderian, 2019). Patents offer unique insights into technology and business activities of firms that could not be assessed by external parties otherwise (e.g., Ernst, 2003;Ernst and Omland, 2011). Patenting firms commonly rely on more than one patent, generating patent portfolios with multiple patents in multiple jurisdictions through their inventive activities (Conley et al., 2013b).
Measuring the impact of research and development or innovation activities proves difficult (Bican and Brem, 2020). Commonly, innovation managers turn to patent analytics (e.g., Pavitt, 1985;Arundel and Kabla, 1998;Hall et al., 2005). For this purpose, various patent indicators have been applied, from patent portfolio sizes and patent citations toward data mined from patents and smart patent indicators (Allison et al., 2004;Buehler et al., 2017;Fankhauser et al. 2018;Guderian, 2019). This variety in patent indicators stems from the skewness of patent values and commercial applicability (e.g., Chesbrough et al., 2006;Gambardella et al., 2008;Webster and Jensen, 2011).

Patent analytics in crises: the COVID-19 pandemic
Patents are cost-intensive over their entire lifetime, as filing and maintenance costs are due in each jurisdiction where patent protection is sought (Duhigg and Lohr, 2012;Cho et al., 2018). Consequently, maintaining existing or filing new patents is challenging when budgets are constrained, as in crisis situations (e.g., Archibugi et al., 2013a; De Rassenfosse and van Pottelsberghe de la Potterie, 2013; Hud and Hussinger, 2015). Here, cost-reduction is crucial when it comes to patents, while simultaneously ensuring the continued protection of relevant innovations and technologies (Helfgott, 1993;Harhoff et al., 2009). This ensures interpreting patents as strategic assets, which allow leveraging existing capabilities and deploying promising business opportunities (Rivette and Kline, 2000;Di Minin and Faems, 2013;Bican et al. 2017). However, this strategic view of intellectual property is underrepresented in the extent literature on innovation management in crises: Only a few early attempts to link propositions to overcome the crisis and intellectual property exist (e.g., Machuca-Martinez et al., 2020;Tietze et al., 2020).
Although scholars have acknowledged that patent information can be used, among others, for technological forecasting, business planning, or strategic 4 R&D Management 2020 R&D planning, the value and use of patent information in crises have largely been overlooked, resulting in a knowledge gap (Campbell, 1983;Ashton and Sen, 1988;Ernst, 2003). By means of this research, we fill this void.
In the current COVID-19 pandemic, this approach implies an understanding of the biology of the peculiarities, root causes, and effects of the virus. Otherwise, strategies to cope with the pandemic become arbitrary, ad hoc, and rely on gut feelings (Müller, 1985;Cooper and Edgett, 2010;Bessant et al., 2015). In the case of the COVID-19 virus, the human body reacts in the following manner: Cytokines are the proteins released as immune response against infections, genetic disorder, or autoimmune diseases (Testar, 2020). They cause localized inflammation in the region of attack (Van Der Meide and Schellekens, 1996). However, in exceptional cases, there is an uncontrolled release of pro-inflammatory cytokines that ultimately cause the body's own immune system to work against itself leading to multiple organ failures and even death (Wang and Ma, 2008). This unrestrained expression of cytokines and other immune response cells and mediators leading to unchecked activation of the immune system is referred to as 'cytokine storm' (Dance, 2020). Cytokine storm has been encountered in patients during the outbreak of SARS-CoV (Huang et al., 2005) and H5N1 influenza virus (Yuen and Wong, 2005). It is also associated with non-infectious diseases like graft-versus-host disease (Ferrara, 1993), multiple sclerosis (Link, 1998), pancreatitis, juvenile arthritis (Goodman, 2020), lupus, and Still's disease.
The COVID-19 virus triggers a similar response once it enters the lungs of the infected patients (Van Der Meide and Schellekens, 1996;Goodman, 2020). When the immune response to this infection becomes uncontrolled it results in uninhibited production of cytokines that cause cell death at the tissue lining the walls of the lungs leading to pneumonia and oxygen shortage. This initiation of lung failure further leads to acute respiratory distress syndrome ARDS (Ye et al., 2020) and is followed by the failure of other organs. Thus, drugs that act as immunosuppressants and thus block the immune cell mediators (cytokines) are being tested to treat the COVID-19 infection. The COVID-19 infection causing SARS-Cov-2 virus belongs to the beta class of the family of coronaviruses termed Coronaviridae like the SARS-CoV and Middle East respiratory syndrome (MERS) virus (MERS-CoV) . These are single-stranded RNA viruses with four kinds of structural proteins (Sidedell et al., 2010;Aronson, 2020). The membranes of these viruses have glycoprotein spikes which are responsible for their 'crown' like or 'corona' like appearance under the microscope (ScienceDirect, 2020;Zhang, 2020). These spike proteins anchor the virus onto the host human cell; this process is the same in SARS-Cov, Mers-CoV, and SARS-Cov-2 viruses . There is more than 70% similarity between the genetic sequence of SARS-Cov-2 virus and the SARS-CoV virus and more than 50% with MERS coronavirus (Park et al., 2020).
There is no known antidote against the SARS-Cov-2 virus. In the current crisis, research in treatment methods for COVID-19 infections is based on the knowledge that is available from the SARS and MERS epidemic in the past and treatment methods so far have been adopted from these past infectious respiratory disease epidemics (Park et al., 2020). One of the treatment methods under study is the development of antibodies that target the Spike protein of the virus . The other method that is pursued is the development of an mRNA vaccine that can trigger the body's immune response to attack the virus (Evans and Longstaff, 2020).

Data
To address the identified gap in the literature, we chose an exploratory, qualitative multiple case study research approach (Yin, 2018). Due to the lack of prior empirical substantiation as well as the nature of our research question ('how'), such qualitative research design was required (Eisenhardt, 1989;Yin, 2018). We rely on four distinct situations of cases that are based on longitudinal and cross-sectional patent data (Yin, 2018); two of which address the downside and two the upside of crisis situations for innovation. We searched cases that address the downside to study how patent analytics can be used to circumvent budget constraints by discerning patents that can be monetized (Case 1: Adidas AG's Patent Families) or abandoned to achieve cost-savings (Case 2: Adidas AG's Annuity Fees), and cases that address the upside to investigate how patent analytics can be used to detect key biotechnology firms (Case 3: CureVac's Patent Portfolio) that are likely to successfully develop treatments and vaccinations (Case 4: COVID-19 Treatment and Vaccination) against COVID-19.
The respective cases were identified in the business press related to the COVID-19 pandemic at the height of the April 2020 lockdown in Europe. Adidas had been in the news for surprisingly obtaining a state-backed loan from the federal German bank KfW, while CureVac drew a prolonged relocation interest by U.S. authorities (e.g., Bennhold and Sanger, 2020;Erhardt, 2020). We obtained the corresponding data, including all measures, from the PatentSight Business Intelligence Analytics Software as of May 7, 2020. For the first two case studies, we queried information on the patent portfolio owned by Adidas AG, including its subsidiaries and subsidiaries' subsidiaries, using the databases' interface. The unit of analysis is the patent family, i.e., the 'set of either patent applications or publications taken in multiple countries to protect a single invention by a common inventor(s) and then patented in more than one country' (European Patent Office, 2017). The Adidas patent portfolio comprises a total of 875 patent families, of which 434 are active. The empirical unit for case study one is the patent family. The empirical unit for case study two is the annuity fees due per patent office and year.
For case study three, we queried information for the patent portfolio owned by CureVac AG, including its subsidiaries and subsidiaries' subsidiaries, using the databases' interface. The unit of analysis is the patent portfolio. The CureVac patent portfolio comprises a total of 123 patent families, of which 105 are active. The empirical unit for case study three is the patent portfolio strength (Ernst and Omland, 2011).
For case study four, we identified patent families related to Coronaviridae viruses. For the patent searches, we use keyword combinations in patents' titles and abstracts as well as Cooperative Patent Classification classes (Swiss Federal Institute of Intellectual Property, 2018; Guderian, 2019). The unit of analysis is the patent family. Cytokine storm treatment, COVID Spike protein treatment, COVID Spike protein vaccine, and COVID mRNA vaccine-related patent search yield totals of 9,494, 271, 115, and 452 patent families, of which 3,059, 128, 48, and 266 patent families are active for these individual technologies on May 7, 2020, respectively. The empirical unit is the number of patent families per owner and their average quality (Ernst and Omland, 2011).

Analyses
Our analyses are conducted in the PatentSight Business Intelligence Analytics Software. For Adidas' patent portfolio, we use the values of the Internal Technology Relevance on the abscissa and the External Technology Relevance on the ordinate (see Appendix for corresponding patent indicator definitions). Each patent family owned by Adidas is plotted according to its remaining lifetime values. For the second case study, we filter for all of Adidas' patent families and the two subsamples derived in the first case study to detect the annuity fees. For the third case study, we filter for CureVac's patent portfolio and display the Patent Asset Index for their technology clusters longitudinally from 2001 to 2019. Moreover, we also filter for the Patent Asset Index of other firms' patent families that build on CureVac's patent families as prior art. In addition, we collect data on protected authorities and inventor locations. For the fourth case study, we search for patent families concerning Coronaviridae viruses and related treatment methods. In light of the crisis, it is necessary to search for patent families and their owners that possess the required preliminary inventive know-how in this field. Our approach in patent searching is to identify these key players who have already performed the groundwork for the fight against the COVID-19 infection in terms of treatment methods such as for cytokine storm. The keywords related to the cytokine storm treatments are searched for in patents' titles and abstracts. Diagnostic methods in this field are not being considered which means that the corresponding CPC class G01N has been excluded. This process ultimately reveals the patents that explicitly disclose inventions concerning treatment methods of cytokine storm. Searching for patents based on the inventions made after the SARS and MERS outbreak can provide the already available knowledge base regarding the ongoing approaches against the virus. The patent search is focused on the inventions in the treatment of SARS or MERS infection targeting the Spike or S protein or the mRNA vaccine. Each of the specific treatment options is investigated more closely by analyzing the patents related to each of the Spike protein-based methods and mRNA-based methods separately. The Spike protein-based keywords are searched in patents' titles and abstracts, while the SARS-and MERS-related keywords are searched in patents' titles, abstracts, and claims. A similar approach for mRNA vaccine-based patents is followed.  Figure 2 depicts the patent office annuity fees of Adidas' patent portfolio due until the end of the patents' lifetimes. While the first row depicts the entire portfolio, the second and third rows depict the annuity fees for the patents that may be abandoned and sold or licensed-out, as identified in the first case study. In total, Adidas has to pay approximately USD 20 million in annuity fees to the respective patent offices like the USPTO and EPO (see Figure 2). For the abandonment candidates, Adidas has to pay about USD 7 million in annuity fees, while for the sale and licensing-out candidates, about USD 4 million have to be paid. The cost-saving potential of USD 11 million accounts for 55% of all annuity fees due.   Figure 3 depicts the Patent Asset Index of CureVac's patent portfolio separated by the technology cluster to which its patent families are classified using semantic rules of similarity (Hartmann, 2020;PatentSight GmbH, 2020) Figure 4 presents the Patent Asset Index of the patent families owned by other firms that cite CureVac's patent families. This allows to identify closely related firms that build on CureVac's patent families as prior art. Among the top 15 firms that build on CureVac's patent portfolio with the strongest patent portfolio, we identified biotechnology firms such as Moderna Therapeutics and Acuitas Therapeutics, but also established pharmaceutical firms such as Roche, universities such as Harvard University and the University of Mainz as well as research institutes like the Ludwig Institute.

Case 3: CureVac's patent portfolio
On a world map, Figure 5 highlights where CureVac's patent families are active and Figure 6 shows where the patent families have been invented  8 R&D Management 2020 based on inventor locations. CureVac's patent families originate in Germany, with few co-origin authorities (e.g., Switzerland, USA). In contrast, CureVac protects its patents across the globe, with the strongest portfolios active in the United States. Figure 7 depicts the top 15 patent owners of Cytokine storm treatment with their respective portfolio sizes, average Competitive Impacts, and Patent Asset Index values. Moreover, besides big pharma (e.g., Bristol-Myers Squibb), small biotech firms (e.g., Ono Pharmaceutical) are active in Cytokine storm treatment. Focusing on COVID Spike protein treatment, Figure 8 illustrates the filing activities on patents for this treatment. Incorporating active, inactive, and pending patents in green, red, and yellow color, the filing activities show peaks for 2003 to 2004 as well as 2013 to 2017, following the SARS and MERS outbreaks, respectively. The most prominent case is Gilead Sciences, whose Remdisivir is the frontrunner in the race toward the treatment of cytokine storm in covid-19 infections (PMGroup, 2020). Aduro Biotech and Ono Pharmaceutical's portfolios feature low numbers of patent families but high average qualities. These firms are known to possess drugs that target inflammatory and autoimmune diseases and have already been taken note of in the current pandemic scenario (Aduro Biotech, 2020;. Regeneron has reported the trial test of its arthritis drug Kevzara to treat cytokine storm in COVID-19 patients   (Erman and Threlfall, 2020). Roche is also conducting tests on its arthritis drug ACTEMRA (DeArment, 2020b;Liu and Miller, 2020) to treat COVID-19 pneumonia, while Incyte has reported its plan for the trial of its blood cancer drug as a potential treatment for COVID-19 cytokine storm (DeArment, 2020a).

Case 4: COVID-19 treatment and vaccination -competitive landscape
For the COVID Spike protein-based search, only 48 active patent families with a Patent Asset Index of 86 are identified. Discerning them by owner type (Figure 9), 64% are owned by firms, 35% by research and governmental bodies, and 1% is co-owned. University of Oxford features among the top patent-owners and the retrieved patents include the one disclosing the specific adenovirus vector ChAdOx1 based on which the current vaccine the whole world is waiting for and which has been  licensed to AstraZeneca, has been developed. In the mRNA-based vaccine-related patents, the two strongest patent portfolios as per the bubble area that indicates portfolio strength, are owned by CureVac and Moderna Therapeutics, where CureVac owns five and Moderna Therapeutics three patent families. The average Competitive Impact values are quite high, but portfolio sizes relatively low with single patent family for firms like Sanofi, while Emory University has six patent families with lower Competitive Impact (see Figure 10).
Latest news states that Moderna's pioneer vaccine technique with its mRNA-1273 in treating COVID-19 infection has got FDA approval to proceed to the second phase of testing (Saplakoglu, 2020). This vaccine is based on the research Moderna was conducting for the MERS vaccine (Harbert, 2020). CureVac has received an 80 million grant from the European Union to develop an mRNA based COVID-19 vaccine (Bahrke and Grammenou, 2020).

Discussion
Innovation management may rely on patent analytics to overcome crisis situations like the current COVID-19 pandemic based on data-driven instead of ad-hoc decisions. As shown in the case study one, patent analytics can be applied to identify patent families that can be abandoned, sold, or licensed-out. Using citation metrics, we show that patent families with below-average internal citations may be considered candidates. Looking at external citations from third-party patent portfolios and considering aboveor below-average values may further support abandonment, sale, or licensing decisions. Adding annuity fee considerations, as shown in case study two, provides the first indication of potential cost-saving opportunities. However, these annuity fees cover patent office fees only. Consequently, associated legal fees for attorneys or patent enforcement costs diminish. Proceedings from the sale of patent families or licensing fees improve the financial position further.
Patent analytics allow for a better understanding of firms' strategies through their patent portfolios, as shown in case study three. For example, firms may identify where to seek protection, where to create R&D centers, where white spots exists, or where to find collaboration partners (e.g., Ernst, 2003;Somaya, 2012;Di Minin and Faems, 2013;Guderian, 2019). Applying patent value indicators to detect firms with above-average patent portfolios related to Coronaviridae virus treatment and vaccination, allows for data-driven innovation management decisions. For example, CureVac's average Competitive Impact (at 18.1) as a measure of average patent quality is about 18 times higher than the average of all active patent families worldwide. Hence, local governments' inclinations to shield these firms from acquisition by foreign entities or relocation through foreign governments, become comprehensible. Also, in terms of the patents' origins and the geographical scope of protection, we see stark differences.
Information mined from patents allows measuring innovation output and future knowledge flows, as shown in case study four (Buehler et al., 2017;Ernst et al., 2020). The innovation-driven patent activity shows a rise after the SARS and the MERS outbreaks in the past. This brings forth the trend of increased innovative activity after pandemics. In addressing crisis situations, the conventional patent searching process where the field of search is kept broad needs to be substituted by more narrow, focused approaches to bring into light the information about those entities as in patent owners that explicitly disclose the R&D and invention in a particular area. Consequently, firms like Moderna and Curevac as well as institutions like the University of Oxford are identified. Innovations are lengthy processes and in their development, they leave behind footprints that can be identified through various methods, one of which is patent analytics.

Theoretical implications
Prior research identified crises to impact innovation in firms (Döner, 2017;Teplykh, 2018;Antonioli and Montresor, in press). A frequent response to crisis-induced detriments is the reduction of innovation activities to cope with issues such as budget constraints (Döner, 2017;Cruz-Castro et al., 2018). Simultaneously, some firms manage to use crises for their benefits to emerge even stronger (Archibugi et al., 2013a;Cefis and Marsili, 2019). While in normal (non-crisis) times innovation management decisions are often crafted and executed using data like patent indicators, patent analytics have so far been largely neglected in responding to the challenges for innovation management imposed by crises. As our key theoretical contribution, we discuss and establish the relation between crises and patents. Thus, as shown in Figure 11, we complement the two dyadic relations between (1) innovation management and crises and (2) innovation management and patents that have been established in prior research by transforming these into a triadic relation between innovation management, crises, and patents.
As visible in the findings from our four case studies, we address firms' internal and external spheres. Patent analytics allow innovation management to craft own or forecast third-parties' probable responses to crises by data instead of ad-hoc or gut-feeling decisions. Moreover, our case studies address multiple levels of analyses, spanning inventions/technologies, firms, and competitive behavior.
Herein, we expand on recent early attempts to link propositions to overcome the crisis and intellectual property (e.g., Machuca-Martinez et al., 2020;Tietze et al., 2020). The findings from the patent analytics presented above confirm Cefis and Marsili (2019) in that technological innovations might trigger firms' survivals in crises: 'R&D investment is shown to be a poor choice for general firms to survive; however, it is an effective strategy for firms that are innovative and capable of producing intellectual properties during recessionary periods' (Jung et al., 2018).

Practical implications
The insights of these studies yield valuable implications for innovation decision makers as well. First, innovation management can realize cost-savings by identifying patent families that do not add substantial value to own or third parties' business strategies. This budget may be used to maintain innovation activities in crises' toughened endowments or circumstances. Second, annuity fees show minimum prices or royalty payments that need to be achieved in sale or licensing negotiations to reach break-even. Third, assessing firms' patent portfolios allows to detect technological and strategic fit between business strategies, innovation capabilities, and patent portfolios. This becomes visible in CureVac's patent portfolio, as almost all of its patent families originate in Germany, whereas broad patent protection has been sought for in many developed and developing countries such as the G-7 and BRIC nations. Fourth, patent information may reveal data-driven predictions for firms that are likely to succeed in developing treatments and vaccinations.

Limitations and future research
As the COVID-19 pandemic is ongoing the results presented can only constitute preliminary assessments. Actual performance effects of management strategies and potential differences to non-data-driven decisions cannot be compared at this stage, requiring additional review once the pandemic has passed. Moreover, it is currently impossible to complement patent data with corporate data or actual decision -making schemes from corporate representatives to identify firms' internal reasonings. However, the data used for the case studies are derived from objective, harmonized, and publicly available information. This opens arrays for future research, particularly related to innovation management in times of crises like the current COVID-19 pandemic via the use of data-driven strategies derived from patent analytics. First, as firms proceed to cope with budget constraints and budget cuts, studying patent families actually sold, licensed, or abandoned, and comparing their patent quality and strength metrics might provide additional insights into value-driven patent management (Wurzer et al., 2016;Weibel and Freytag, 2019). Particular focus may be set on assessing quality differences for innovations generated in-house versus those externally acquired. Second, future research may move from qualitative analyses and case studies toward empirical analyses to study actual performance effects of patent data-driven decisions in crisis situations. A promising approach in this context may be to expand the existing models toward multilevel or hierarchical models to combine patent data on individual patents or patent portfolios with financial, economic, or medical data (e.g., Raudenbush and Byrk, 2002;Tabachnik and Fidell, 2018). Third, research may shift to longitudinal considerations, which is already partly performed in the case studies presented in this study. Looking at patent data trends and development patterns over time may yield additional insights into how firms overcame the crisis, allowing to make derivations for future pandemics or similar situations to come (e.g., Grover, 1999). This could support the detection of firms' development of Coronaviridae treatment and vaccination development to end the pandemic.