17 February 2020
The interactive component of this report is best viewed on desktop.
Social networks enable innovation. However, social network analysis of Australia’s innovative efforts has thus far been limited.
This research, a joint project between the United States Studies Centre (USSC) and LinkedIn, is one attempt at tackling that challenge. For this project, we looked at LinkedIn connections between start-ups, start-up founders, and venture capital firms in Australia, New Zealand and the United States as proxies for social network connections. We limited our focus to agricultural technology (AgTech) networks, building on prior seminal research from USSC on Australia’s AgTech industry evolution and benchmarking.
We are proud to present one of the first in-depth analyses of a relational network in the VC-entrepreneur economy using the rapidly growing AgTech market as an exemplar.
While deliberately limited in scope, the data is striking. Australian AgTech networks are not only behind the United States in cohesion and interconnectedness, but they also trail New Zealand in cohesiveness as well as comparative connections to the United States. Australia’s near neighbour has a fifth of Australia’s population yet is somehow performing better in building ties to the world’s largest AgTech market.
To address Australia’s AgTech challenges, we conclude this report with forward-thinking policy recommendations.
We thank LinkedIn, particularly the talented Joel Van Veluwen, for the many hours put into making this research possible.
Social networks are an integral component of entrepreneurial ecosystems.
From Silicon Valley to Israel, dense, globally-minded networks have facilitated the movement of talent, attracted funding, directly affected collaboration and, ultimately, facilitated innovation. Multiple studies indicate shared cultural and physical ties between geographies are key to building some innovation ecosystems.[^1] There is a growing recognition that intra-network density is not the only indicator of innovation success. There is strong evidence to suggest that the best innovation occurs within dense networks which contain actors that are open and receptive to external and novel knowledge flows. The inverse appears not to be the case. It has long been held that networks which are geographically removed from industry-relevant clusters have significantly less innovative output, even when they are located in manufacturing hubs.[^2]
This study builds on an academic body of research into social networks between centres of innovation capital and geographically distant entrepreneurial ecosystems seeking funding. It aims to analyse this connection and ask the following questions: How developed are Australia’s networks in particular innovative sectors? Are Australian networks well connected, or are Australia’s entrepreneurs working in silos? And ultimately, how can policy help build Australian networks further?
According to our analysis of proprietary LinkedIn data on one of Australia’s burgeoning innovative sectors, AgTech (where agriculture meets technology), Australia’s networks are less cohesive and interconnected than US AgTech networks. In comparison, New Zealand’s AgTech networks appear to be smaller yet more cohesive, interconnected and — critically — more connected to the United States AgTech networks. As the largest AgTech market in the world — estimated to be valued at US$10.2 billion,[^3] and accounting for roughly 65 per cent of global AgTech investment[^4] — connections to the United States are vital but Australia is clearly missing out.
Social network analysis is exactly what it sounds like: the analysis of social networks. Used for everything from identifying potential terrorists to determining “influencers”, social network analysis is a sociological method.
The method defines social networks in terms of nodes (individual actors, people, or entities within the network) and the ties (relationships, links or interactions) between them. The nodes in our networks are AgTech companies and funders. The ties between them are established by considering the LinkedIn ‘connections’ between the individual sub-nodes of the network: the employees, board members and founders of those companies and funders.
In terms of describing the nodes and ties on the social network maps, there are two primary variables: the centrality of nodes and the density of the ties.
The centrality of nodes indicates the importance of nodes in a network. Stanford University, for example, would likely be the most central node of any social network analysis of Silicon Valley because so many key individuals in Silicon Valley either studied, worked, or lived on Stanford’s Palo Alto campus.
The density of ties is the number of actual connections divided by the maximum potential number of connections. This ratio gives an indication of network cohesiveness. The denser the network, the closer the output of that ratio will be to the number one.
For example, consider a network with six organisations and 15 links between them (Figure 1). In this simulated example, the maximum potential number of links is 15. A graph with a link density of one is also known as a complete graph. In other words, everyone is connected to everyone. As the link density decreases, the cohesiveness of the network declines. Take the same six companies, now with seven links. The link density is now 0.5. Evidently, there are far fewer links between companies. If the companies had no connections, there be a link density of 0 and they would no longer be considered part of a network.
Figure 1: Density of network
This simulated example of a social ecosystem consisting of a set of actors and the existing ties between them could also be a representation of a venture capital social network. In this case, the presence of a tie between any two venture capital firms indicates they have at least one investee company in common: that is, that both VC firms cooperated in the financing of the same company or companies. That connection is indicated in the figure by a tie between two dots.
One previous example of social network analysis studied the network map of Silicon Valley during the 1995-1998 period in this way.[^6] There were 111 firms and 312 links between the firms. The 312 connections represented any time a connection was forged, not the number of total times a co-investment was made. The resulting link density of 0.05 is low — quite a few VC firms had not co-invested in a company over the three years.[^7] Yet there was a subset of the network that was highly dense — they were the important, key players in the network.
This framework measures the importance of individual firms by looking at the number of connections the firm has to other firms that are highly connected in the network. A Silicon Valley firm may have been connected to many other Silicon Valley firms, for example, but if these were isolated connections then they did not matter. Ultimately, start-ups want to be connected to funders who are likely connected to a large number of firms and other potential funders in the network. Social network analysis offers one a measurement of a firm’s success in making such connections.
To conduct our social network analysis, USSC gave LinkedIn a list of key individuals who founded, funded or led AgTech start-ups in the United States, Australia and New Zealand. To give an indication of the strength of the US network vis-à-vis the other AgTech networks, we analysed the number of connections between the individuals within each self-selected start-up and funding entity.
Our conservative selection process was as follows: We chose every AgTech entity we could confirm was both AgTech focused and active as of 2019 in the three countries. Some definitions of AgTech, also referred to as ‘Agrifood Tech’, include technologies relating to consumer-facing components of the agricultural supply chain, such as restaurant and retail innovations — think e-commerce-enabled meal kits such as Blue Apron or HelloFresh. The criteria we used to select firms do not include these sorts of technologies.
Using this definition, USSC used its prior research into AgTech to make lists of active AgTech funders and start-ups in the United States, Australia and New Zealand. It then gave these lists to LinkedIn for it to mine its proprietary data for connections between the lists. At no point did LinkedIn give the United States Studies Centre any private data of LinkedIn users — all data was grouped together by company.
This selection process found a total of 84 AgTech funding and start-up organisations in Australia, 49 such organisations in New Zealand, and 124 in the United States. The AgTech employees and board members of either start-ups or funders totalled 631 in Australia, 609 in New Zealand and 2,467 in the United States. Our analysis encompasses the employees and board members of each funder and start-up — although it should be noted that the data that LinkedIn gave to USSC did not give the connections of each individual.
There is, of course, the potential that one company could have a ‘connection’ with another in the absence of a LinkedIn connection. Attempting to prove the negative in this case is prohibitive when analysing three geographically distinct and dynamic networks.
The inverse is also true. A LinkedIn connection between two individuals of two AgTech firms does not infer that those two firms, or even those two individuals, will ever collaborate in a meaningful way.[^8]
When taken in isolation, the data presented here is limited by these and a series of other external factors, including, for example, whether individuals update their profiles, or whether there are individuals who do not use the platform at all. But as LinkedIn is the most-used online professional networking platform, and AgTech is a highly digitally-literate industry, we see it fitting to use the data as representative and as another block in our ongoing research into innovation and networking in the Australian AgTech sector. Our previous findings regarding the flows of venture capital and attitudes towards innovation allow us to bridge some of these methodological gaps and to use this data as a soft measure of networking proactivity in and across the three markets.
The US AgTech network is more than three times as interconnected as the Australian network, with AgTech start-ups in Australia having link density of 0.11 compared to a 0.38 link density for their US counterparts.
Since links have a distinct connection — whether directly in joint ventures or peripherally by sharing contacts, link density can also be explained as a measure of the cohesion of the network. The sheer difference in scale between the United States and Australia is made most clear when considering the median number of connections per Australian start-up is 11 while the median number of connections per US start-up is 69.
Figure 2: AgTech in-country networks on LinkedIn
Table 1: Density of connections within US, Australian and New Zealand AgTech start-ups and funders
Australia’s AgTech network is not only less connected than the United States; compared to New Zealand, Australia is well behind. For New Zealand start-ups, the median number of links per firm is 58 while the link density is 0.48 — more than double that of the link density in the United States. Start-ups in New Zealand are tightly connected — likely a factor of geographic isolation and a small population — whereas US and Australian networks are larger but less densely connected. Large networks are often more likely to be sparse as the number of possible links in a network expands by the total number of links already in that network each time a new node is added.
For funders, the median number of links in New Zealand is on par with the United States — in fact, New Zealand does slightly better than Australia regarding connectivity to US venture capital on LinkedIn. Australia lags, but this gap is not as large as the distance between the number of links with American start-ups vis-à-vis Australia.
Table 2: Density of connections in AgTech networks between US and Australian entities, and US and New Zealand entities
Figure 3: Australian start-ups have fewer connections on average compared to all other firms in the network
Beyond the broad country comparisons, it is important to understand what this analysis looks like in a tangible sense. The clearest way to do that is to determine the most centrally located firms in each country.
Social network analysis has a tool for this, known as eigenvector centrality. Ultimately, the eigenvector centrality is an equation that measures the influence of a node in a network — the higher the eigenvector centrality, the better connected and more influential the node.[^9]
A good way to think of a node with a high eigenvector centrality would be to consider the families in a rural community and their connection to a church. While many families in the area may not have direct links to each other, as churchgoers, they have links to the priest who is considered an influential authority figure. Unlike these families, the priest is also connected to other priests in other communities. Therefore that priest would have a much higher eigenvector centrality score in the network. Families with stronger links to that priest would have higher eigenvector centrality scores than those that don’t.
We’ve used a complex equation using linear algebra and matrix theory (see appendix below) to rank the best-connected entities in the AgTech networks of the United States, Australia and New Zealand. The highest possible eigenvector centrality score close to 1.000 indicates a node that is immensely influential within a network. A node with an eigenvector score close to 0.000 is considered to have very little influence within that network.[^10]
Table 3 looks at the important firms within the combined Australia-New Zealand-US AgTech network on LinkedIn. In this we are seeking to measure how Australian firms are connected to key players in the United States compared to New Zealand firms. It is important to note that as the US market is the largest and most influential, firms — both funders and start-ups — located within it are more likely to be connected to other important firms and more likely to have high eigenvector centrality scores. Australian firms and New Zealand firms have fewer connections to important firms and thus score lower. Note, however, that a firm in Australia that has a low eigenvector centrality score in a set including the United States and New Zealand can be highly connected in its own domestic network. A firm that is connected to other firms in the local network might become less well connected when factoring in the three-country network in the equation.
The best connected AgTech start-up in the United States is FullHarvest, a San Francisco-based online marketplace for ‘ugly’ or surplus produce. Founded in 2015, it endeavours to connect growers to food and beverage companies in order to reduce massive amounts of waste in the supply chain. Full Harvest has raised a total of US$11.5 million in funding over three funding rounds from sources that include US funders Wireframe Ventures and Spark Capital.[^11]
In Australia, the best-connected firm is FluroSat, an agricultural data analytics firm that develops software to help farmers use data, agricultural science, remote sensing and artificial intelligence to monitor all aspects of their practice. Headquartered in Sydney with offices in Canberra, San Francisco and Kyiv, FluroSat was founded in 2016. Since then, it has raised a total of US$8.6 million in funding over five rounds from Australian funders including MainSequence and AirTree Ventures as well as Microsoft’s Venture Fund.[^12]
And in New Zealand, the best-connected firm is BioLumic, a Wellington-based start-up founded in 2012 that uses applied agricultural science to increase crop yields with ultraviolet devices. BioLuminic raised US$14.6 million in funding over six rounds including investment from US funders Finistere and Radicle Growth as well as Canadian funder Canopy Rovers.[^13]
When looking at the eigenvector centrality of the firms in each country’s AgTech network, the United States clearly dominates, with US firms significantly better connected than firms in Australia and New Zealand. Australia is even worse than New Zealand on this metric. The difference in eigenvector centrality between, respectively, the highest-ranking firm in New Zealand (BioLumic) and Australia (FluroSat) is 0.094 — a fairly significant amount. Also significant is the gap in the average eigenvector centrality of Australia’s top 10 most connected firms versus their New Zealand counterparts: 0.063.
Table 3: Which firms are the best connected to US firms? AgTech start-ups and funders ranked
Figure 4: Network link distance from any firm to an important firm
Research has firmly established that firms in well-connected clusters have access to more information and are more knowledgeable about the industry.[^15]
In addition to being one of the largest and most significant AgTech hubs in the world, Silicon Valley is also one of the best-connected AgTech hubs. A simple glance at the San Francisco AgTech network compared to the AgTech networks of Auckland, Brisbane, Melbourne and Sydney confirms as much: San Francisco’s AgTech network is undeniably denser.
Sydney has a smaller AgTech ecosystem than Silicon Valley, though the difference in connectivity is still not commensurate. Sydney’s network — given its smaller size — is denser (0.51) than Silicon Valley (0.47). But Sydney appears to have fewer outliers in the geographic cluster than Silicon Valley.
Australian AgTech networks in Melbourne and Brisbane are considerably smaller, less developed, and less connected than in Sydney. Australian AgTech networks outside of Sydney have a lot of room to grow.
Figure 5: Geographic clusters
Figure 6: LinkedIn connections from Australian AgTech start-up founders to US or Australian funders over time
The United States is a destination market for many of the world’s most ambitious start-ups. It is also a key potential source of funding. Neither Australian nor New Zealand AgTech networks have particularly close ties with the United States. With that said, New Zealand’s comparatively smaller AgTech network has denser ties with the US AgTech network than Australia’s AgTech network.
The fact that many Australian AgTech firms do not have links to Silicon Valley firms not only means that start-ups risk missing out on institutional knowledge from the movement of talented individuals throughout the network, but they also risk missing out on interest from investors — many of whom have technical backgrounds and networks in the field. It also means Australian firms may miss out on opportunities to expand into the US market. Prior USSC research found that Australia’s venture capital market invests dramatically less in AgTech per capita than most developed nations.[^16] An underdeveloped network is less likely to support the opportunity for investment from US-based venture capital firms to bridge the gap between early-stage and late-stage funding.
Recently, Harvard University showed that in early-stage venture capital investment trust relationships have outsized importance in the decision making process.[^17] Moreover, anecdotally it has been reported that less than one per cent of Silicon Valley deal-making is done via a “cold” introduction. This means that if a start-up management team is not known by a lead investor at a particular VC fund then it needs to be connected to someone who can make an introduction to have the best chance of an investment process being undertaken. In this way, networks and relationships are a gating component for start-ups for VC investment. Consequently, the findings we report here, in conjunction with the previous work from the USSC,[^18] indicate that Australia’s AgTech ecosystem is facing significant structural challenges to success. More broadly these findings also suggest that a lack of global networks at the firm level may decrease Australia’s competitiveness for venture capital-based foreign direct investment.
Australia and New Zealand both have strong global and domestic conditions for AgTech innovation. They have strong agricultural reputations, world-class research facilities, free trade agreements with much of Asia, years of acting as a testbed for new technologies, strong intellectual property rights, and relatively market-friendly regulations. The most distinct difference between the two nations is that from the size of its population to capital markets, New Zealand is smaller than Australia by practically every measure.
Yet despite — or perhaps because of — such differences in size, New Zealand’s AgTech network appears further along in its development. The New Zealand AgTech network is more cohesive, more interconnected, and has as many total connections to the US networks as the Australian AgTech network.
The perception among AgTech industry experts is that New Zealand is “years ahead” of Australia and the network data appears to confirm it.
Beyond network data and anecdotes, there are tangible ways in which New Zealand has been comparatively more successful than Australia in both its development of AgTech as well as its development of an innovative economy:
So what is New Zealand doing right? On the whole, there are two general themes: widespread government encouragement and a global mindset. The results of this are quite evident.
Agriculture has an outsized role in New Zealand’s economy. As much as the Australian farm has an iconic role in the country’s economy and history, the agriculture sector is only responsible for 2.7 per cent of gross domestic product compared to 5 per cent of New Zealand’s GDP.[^23] Furthermore, only 14 per cent of Australian exports are agricultural compared to more than half in New Zealand.[^24]
With agriculture’s contribution to GDP more than twice as large in New Zealand as it is in Australia, the stakes for success in AgTech are, as a result, more than twice as high in New Zealand.
The notion that New Zealand needs to “get this right”, could not be felt any more urgently. The New Zealand government has recognised this and responded accordingly, with a whole-of-government prioritising of AgTech, as evidenced by its AgTech industry body, Agritech New Zealand.[^25] In addition to representing the AgTech community, the body has also been actively supported by senior government ministers and has attracted private sector partners.[^26]
According to Peter Wren-Hilton, executive director of Agritech New Zealand, New Zealand AgTech innovations have penetrated global markets for some years now for a simple reason: most of New Zealand’s AgTech ventures have an eye towards international markets.[^27] Like much of the rest of New Zealand’s entrepreneurs, New Zealand AgTech ventures do not dream of merely saturating their domestic market — they are globally-oriented from day one.
The increasing role of technology in New Zealand's exports is symptomatic of this global mindset. Now New Zealand's third-largest export sector, technology grew by A$1 billion in 2018.[^28] In Australia, on the other hand, technology is barely a top five export.[^29] As a percentage of total exports, Australia’s internet and communications-related service exports and intellectual property receipts, are both behind New Zealand’s.[^30]
On a practical level, a global focus means New Zealand AgTech firms are not planning to replicate the success of certain technologies that already exist elsewhere but, compared to Australia, focus more on innovating new and disruptive technologies.
This also means that New Zealand actively seeks out the international AgTech community: discussions with AgTech experts made clear that foreign investors in AgTech feel more welcomed in New Zealand than they do in Australia.
The findings in this report, although based on a limited measure, build on our previous research into innovation and AgTech in Australia. A lack of connectivity to the US market adversely impacts Australia in an increasingly digital global world. In AgTech, where Australia has excellent market potential due to a long history in agriculture coupled with strong agricultural science and a well-educated technology workforce, to be lagging New Zealand should be a wake-up call to policymakers.
Australia can both learn from and work with New Zealand to more effectively seek better connectivity with and investment from the Silicon Valley AgTech network. Initiatives that should be considered by the Australian government include:
More cohesive and coherent advocacy for Australian AgTech
Australia and New Zealand can both expand efforts to leverage their strengths in AgTech in concert. Already a long distance away from most other markets, the two countries are more likely to draw attention when considered together rather than distinct from each other. A coordinating body, the Australia-New Zealand Agritech Council, is in the early stages of development and should be launched and supported by both governments.
Such a body could help build the networks of Australian and New Zealand AgTech in a way that facilitates global growth.
Stronger engagement of venture capital as a key component of building networks and growth
As documented by prior USSC research, Australia is in short supply of the sort of venture capital that is necessary to scale globally.[^31] Australian policymakers should look to incentivise sophisticated AgTech investors from overseas, particularly from the United States, to access world-leading expertise in this sector and to take advantage of the major source of capital.
Foreign venture capital firms supply more than money, they can provide the global networks and expertise in how to scale. Ranging from financial incentives to government-sponsored trips, the Australian government can learn from the efforts of other nations like Israel and New Zealand in attracting and welcoming foreign venture capital firms.
Greater support for AgTech from the Australian government
Particularly following the 2019 release of a shared Australian vision for agricultural innovation and the launch of Austrade’s Agriculture 4.0 Initiative — the initiative that focuses on facilitating foreign investment, exports and collaboration in both AgTech and food tech — the government has accomplished a lot on AgTech.[^32] These efforts support the existing landing pads, Austrade-supported offices that provide market-ready Australian start-ups and scaleups with access to the innovative cities of San Francisco, Tel Aviv, Shanghai, Berlin and Singapore. These have been found to be integral points of contact for Australian AgTech firms looking to go global. Austrade also has two investment specialists in agribusiness and food in Australia and helped facilitate trips by Australian AgTech leaders to the United States. Yet the Australian government can do more.
Austrade can follow New Zealand Trade and Enterprise’s lead and consider subsidising the full costs of trips by AgTech funders from the United States. While facilitating and supporting such trips is key, a significant hurdle is removed if qualified AgTech funders from the United States are incentivised with trips to Australia. Optimally, such trips could be coordinated together by both Austrade and New Zealand Trade and Enterprise so as to reduce cost and increase interest.
A wider understanding of the role of AgTech
The Australian government needs to adopt an understanding of AgTech that goes beyond simply increasing crop yields and efficiencies. AgTech is part of the knowledge economy and is, therefore, not limited by Australia’s crop or land sizes — the whole world is the market. While innovation has become a priority in certain regions across Australia, few government officials are welcoming AgTech into these efforts because it is pigeon-holed as primary industry.
To better appreciate the economic potential of AgTech beyond merely increasing domestic agricultural production, the Australian government should consider supporting a study that would quantify the potential economic impact of AgTech as part of a wider perspective on the knowledge economy. Putting a figure on the potential GDP contribution of the knowledge economy and its component parts, including AgTech, would help both policymakers and the Australian public to better understand AgTech’s unique potential in Australia. The Department of Industry’s new Digital Economy and Technology division is the optimal candidate for leading such efforts.