Platform & Ecosystem Analysis

When the Best Product Is Not Enough

A satellite operator with superior resolution, better revisit times, and lower unit costs nevertheless loses customers to a marketplace that resells competitors’ imagery alongside its own. A ground-station provider with the largest global footprint watches a software layer from an adjacent industry absorb its customer relationships and relegate it to back-end infrastructure. A launch broker with unmatched access to upper stages finds that the decisive price negotiation no longer happens on its platform, but in an aggregator that the broker’s customers trust more than the broker itself.

These outcomes are puzzling only if one reads them through the logic of product competition, in which the better offering wins. They become legible the moment one reads them through the logic of platforms, in which the critical question is not which offering is superior but which interface connects the largest number of counterparties at the lowest coordination cost. The space sector has been slow to take this shift seriously, partly because the dominant firms in the industry still think of themselves as producers, not orchestrators, and partly because platform economics was developed for consumer-internet markets whose network effects were so loud that the theory seemed exotic to an industry built on bespoke hardware and sovereign contracts.

It is not exotic. It is arriving, unevenly but unmistakably, in satellite data, in ground-station-as-a-service, in rideshare launch aggregation, and in cloud-adjacent downstream services. An analyst who cannot read platform dynamics will routinely misidentify the competitor, the threat, and the defensible position.

Where the Ideas Came From

The intellectual scaffolding is recent by the standards of strategic analysis. Jean-Charles Rochet and Jean Tirole’s 2003 paper on two-sided markets gave the field its formal economic foundation: platforms are markets whose pricing and design decisions must account for the fact that the value each side gets from the platform depends on how many users are on the other side. The classic examples — credit cards connecting merchants and cardholders, operating systems connecting developers and users — share a structure that single-sided industries do not.

Annabelle Gawer and Michael Cusumano, first in 2002 and then more systematically in 2014, translated the economics into a managerial language. They distinguished innovation platforms from transaction platforms and wrote about platform leadership as a distinct strategic discipline. Marco Iansiti and Roy Levien, in a 2004 book, gave the field its ecological vocabulary — keystones, dominators, niche players — which made it possible to talk about the health of an ecosystem rather than the performance of a firm. By 2016, Geoffrey Parker, Marshall Van Alstyne and Sangeet Paul Choudary had consolidated this literature into a popular book that placed the platform concept in the hands of a wider strategic audience and coined the idea of “envelopment” — the move by which a platform from an adjacent market swallows a neighbouring one by extending its existing user base into it.

This lineage matters because it explains a recurring confusion. Platform theory is not a single theory. It is an economic claim (two-sided markets exist and behave distinctively), a managerial claim (orchestrating a platform is different from running a pipeline firm), and an ecological claim (ecosystems have health independent of any single firm’s profit). Each claim reaches the space sector at a different speed and demands a different analytical response. The method’s job is to hold all three in view at once.

The Characteristic Move

What platform analysis sees that other methods do not is the multi-sided interaction. A business model canvas describes one firm’s value proposition; a Five Forces reading describes industry structure around a product category; a supply-chain analysis traces inputs into outputs. Platform analysis reframes the unit of study. It stops looking at the firm and starts looking at the interaction the firm facilitates — between data providers and data consumers, between launch customers and launch suppliers, between payload operators and ground stations.

Once the interaction becomes the unit, a series of distinctive diagnostic questions follow. Who are the sides? What cross-side dynamics connect them — does growth on one side pull the other? Are same-side effects positive (more users attract more users) or negative (crowding, quality dilution, price competition among suppliers on the platform)? Can participants multi-home — that is, stay on competing platforms simultaneously — or are there switching costs that lock them in? Where is value created, and where is it captured? These questions have no analogue in single-sided analysis.

The method’s second distinctive move is ecosystem classification. Iansiti and Levien’s taxonomy separates ecosystem participants by the role they play in shaping collective health. A keystone sits at the centre of the ecosystem but extracts moderately, leaving enough margin for complementors and niche players to thrive; a dominator extracts aggressively, hollowing the periphery and eventually weakening the ecosystem that supports it; niche players specialise narrowly and depend on the keystone’s governance for survival. Reading these roles is not the same as reading competition. A firm may be highly profitable and simultaneously an ecosystem dominator whose behaviour predicts the decline of the very platform it dominates.

The third distinctive move is envelopment analysis. Traditional industry analysis assumes that competitive threat comes from within the industry. Platform analysis forces attention onto competitors who are not, technically, in the same market at all — a cloud provider that absorbs a satellite data marketplace by bundling imagery into its broader analytics stack; a geospatial software firm that extends its user base into ground-station bookings; a telecom platform that annexes launch scheduling through adjacency. Envelopment is the platform-native threat, and it is invisible to methods that define markets by product category rather than by user overlap.

The practitioner who internalises these three moves — the multi-sided reframing, the ecosystem roles, the envelopment lens — will see structures in the space industry that a product-centric analyst cannot.

The Method at Work: A Satellite Data Marketplace

Consider a satellite data marketplace in the middle stage of its development. On one side sits a heterogeneous population of constellation operators — some owning optical, some synthetic aperture radar, some hyperspectral assets — who contribute imagery to the marketplace in exchange for access to demand they could not reach alone. On the other side sits a fragmented population of buyers: government analytical units procuring sensor-agnostic coverage, commercial analytics firms assembling multi-source pipelines, and a thinner tier of enterprise users buying spot imagery for specific workflows. The marketplace operator takes a transaction percentage and provides a standardised API, search tools, and a common licence template.

The platform map surfaces a cross-side network effect of moderate strength. More operators on the supply side increase the breadth and price competitiveness of imagery available to buyers, which attracts more buyers, which attracts more operators seeking monetisation for excess capacity. This is the textbook dynamic, and it is real. But the map also surfaces a negative same-side effect on the supply side: operators compete on price within the platform, commoditising their own output. An operator who contributes unique imagery is rewarded by the platform, but an operator who contributes substitutable imagery is punished by it.

Value creation is easy to trace: the platform reduces search costs, standardises licensing, and aggregates fragmented demand. Value capture is more revealing. The largest share of captured value flows, in the early stage, to the platform operator, which sits astride the interaction and controls the API that buyers have integrated into their workflows. The second largest share flows to operators whose data is both unique and in demand — hyperspectral providers, for instance, who have no close substitutes. The smallest share flows to operators whose output is commoditised by the platform’s own matching function.

The ecosystem role reading is sharper. The platform operator, in this example, is beginning to behave like a dominator rather than a keystone. It extracts an increasing transaction share, publishes terms that favour its own analytics layer over independent complementors, and uses the aggregated data to train its own models. This behaviour predicts, not immediate collapse, but a slow weakening of the supply side as the most valuable operators either vertically integrate their own sales channels or multi-home onto competing marketplaces. Ecosystem health is declining beneath apparent platform profitability.

The envelopment risk completes the picture. A cloud provider whose platform already hosts most of the buyers’ analytical workloads can, with modest investment, launch a data-bundling service that uses its existing user relationships to extend into imagery distribution. The cloud provider does not need to build a marketplace from scratch; it needs only to become the default buyer interface and relegate the original marketplace to an upstream supplier. The envelopment threat, in this example, is the structural vulnerability the analysis surfaces — and it is a vulnerability no amount of product improvement inside the marketplace will address.

The analytical finding a practitioner leaves with is not “this platform is profitable” but “this platform is dominator-drifting under envelopment threat, and the trajectory is hidden by current transaction growth.” That is a finding product-level analysis could not reach.

Where It Holds, Where It Zoppica

Platform analysis handles better than any rival the cases where value flows through interaction rather than production. It catches the competitive threats that industry-boundary methods cannot see and it distinguishes healthy ecosystem roles from extractive ones. For sectors where platforms are forming — satellite data, ground-station-as-a-service, launch aggregation, certain classes of downstream analytics — it is the method of first resort.

Its limits are equally sharp. The theory was developed for consumer internet markets where user bases run into tens of millions and the network effects are loud enough to be measured from the outside. Many space-sector “platforms” operate with hundreds of sophisticated participants, not millions. The classical assumption that cross-side effects will dominate is not automatic; at small scale, direct negotiation and relational trust can outperform platform matching. The analyst who applies the framework mechanically will overstate network effects and underweight the fact that space procurement remains heavily bilateral.

The method is also weak on the government variable. Standard platform theory assumes open markets; in space, the largest buyer is often a sovereign procurement agency that ignores platform logic entirely, awarding contracts on strategic rather than transactional grounds. Government as simultaneous anchor customer, regulator, competitor, and occasional supplier is a dynamic the theory does not model well. For regulatory dimensions, regulatory impact analysis is the necessary complement; for the buyer-power and entry-barrier dimensions, a Five Forces reading overlaid on the platform map provides the corrective.

Finally, winner-take-all predictions are the method’s seductive failure mode. The theoretical conditions — strong cross-side effects, high switching costs, expensive multi-homing — are rarely satisfied all at once in space markets. Most platforms in the sector will settle into oligopolistic or differentiated-niche structures, not monopolies. An analyst who predicts concentration should test against historical analogues in adjacent industries before committing to the forecast.

The method pairs well with business model analysis (which describes the individual firm within the ecosystem), supply-chain dependency analysis (which surfaces API and data-format lock-in), and regulatory impact analysis (which addresses antitrust and data-governance exposure that platform governance choices invite).

A Note for the Practitioner