The Wall Street Journal notes that Amazon now allows consumers to purchase and send movies directly to Tivo. This is great, but it does not seem suitably different and maybe a bit less convenient than the movies on demand services already offered via cable. That said, I look forward to the day when Amazon allows me to purchase and send a movie to Tivo as a gift. In that case, I'd send cheap movie shorts (ideally with an attached video message that I would store at Amazon) as a new form of high quality video e-card. I am a little bit surprised that Youtube is not offering video e-cards in similar fashion (through email rather than to Tivo)
Update: well, I was kinda wrong
Update 2: Something closer to a video e-card offering
===================================================================================
The Immigrant Technologist points to an interesting working paper written by several researchers within the HBS Entrepreneurial Management unit. The paper, "Platform Envelopment" takes a look at competition against entrenched platforms with large networks- from digital music players to online recruiters and shipping providers- where competitors may either compliment each other or share some functionality.
The researchers use a variety of case examples to generate three competitive approaches as follows:
A "conglomeration attack" where a company bundles together platforms that share a common user base but provide different services to those users. Examples includeDoCoMo offering payment platforms to mobile users who presumably have credit card accounts and phone companies offering bundled phone, cable, andinternet services
An "intermodal attack" where two companies offering similar (but not identical) platforms begin offering similar services. Examples include DHL offering overnight services in response to new FedEx offerings and Monster.com offering social network-style profiles as Linkedin launched a jobs board
A "foreclosure attack" where a firm begins offering a substitute service in order to place pressure on a complementary service offered by a competitor. Examples include Microsoft offering a "save as .pdf" option with Office 2007 in order to put pressure on Adobe and EBay offering Billpoint (until it failed) to put pressure on paypal.
The writers go on to ask about the factors that may impact success within each of these competitive approaches in the context of user valuations, asking whether a user will see more value in using two or more goods/ services together rather than separately. They suggest that this is a matter of relative valuations suggesting that two platforms should be bundled (conglomeration attack) together when the user valuations of one or both goods and services is substantially higher than the marginal cost of said goods and services at the profit maximizing price.
This suggests to me that the strategies are oriented toward a base of power users, if only because the "average user" should have a valuation closely linked with the price point at which the marginal cost and marginal revenue curves meet (ie, the profit maximizing price). This further suggests something that should be somewhat obvious: convergent services succeed when the user valuation of the sum of the products exceeds the summed valuation of each of the discrete parts. In statistics, we might identify these cases via a test for interaction (*). The paper (because it is a working paper) does not verify this assertion by defining the users (from power users to casual users) who will see the value in bundling.
"Aha!" I say. Consumer valuation is closely related to the "fit" between a particular solution and a job to be done. What if I were to re-frame this in terms of jobs to be done?
Conglomeration Attack: Here we have two functionally independent technologies. Users have adopted one technology to address one set of jobs and happily use a second platform to address a different set of jobs. In theDoCoMo case, I use a mobile phone to keep in contact with other people and I use a credit card to make purchases. In order for bundling to work in this case, the bundled platform will need to satisfy a new, previously unsatisfied job that is important to some consumer subset. It will not be enough to simply treat a phone as an extra credit card. My credit card jobs are suitably satisfied. It should not surprise me that bundling efforts often fail until they move away from consumers who already use both platforms and toward consumers who have access to one platform but not both. Phones become a channel formicropayments in cash based economies with low credit card penetration. To place this in the context of valuation and marginal cost, the MC/V of the bundle increases over the MC/V for each of the constituents because the dominant platform in the bundle (here, the phone) sharply lowers the marginal cost of access to the new component (purchasing) so that the value of being able to make a purchase without cash can suddenly exceed the cost of obtaining access to small amounts of credit.
Intermodal Attack: Power users are likely to use both competing systems concurrently, so switching costs will be low. A DHL power user will also have an online account with FedEx. The most demanding Monster.com users will also have LinkedIn accounts. These users will typically employ both services to address a focal job, whether "I want to reduce my worry when delivering critical packages" or "I want to be sure that I am not missing out on a new and interesting career opportunity." The services are left to compete in their ability to address these core jobs (actually, the jobs/ restrictiontradeoffs). It makes sense, then, that intermodal attacks often result in mirrored behavior. An improved strategy in this arena might result in designing new servcies that really fit consumer jobs to be done instead of just mirroring competing services.
That said, the LinkedIn jobs posting could really be recast as a conglomeration attack while the Monster social network could really be recast as a foreclosure attack. All of these approaches exist in a spectrum. UnlikeDHL/ Fedex, which are conceptually "closer," LinkedIn & Monster approach career networks from very different angles and this creates an asymmetry of switching costs. I imagine that it is more of a pain for aLinkedIn user to recreate a social network within Monster than it is for a Monster user to repost information in a larval profile on LinkedIn. At the same time, it is easier for a LinkedIn user to employ the LinkedIn job search service even while using the Monster.com service in parallel. This creates a market that is advantageous to LinkedIn in the long run even if LinkedIn faces real challenges in replicating Monster.com's voluminous jobs database and even if it is a pain for employers to repost inquiries on LinkedIn. In this case, the jobs are similar but the "ease of adoption" restrictions drive meaningful differences between the services.
Foreclosure Attack: Consumers often employ multiple goods and services in multistep processes in order to address a given job. If I want to get a large couch back from a retail store, I might employ a rental service, a van, a gas station and a dolly. I am often aware of the total cost of this "supply chain" and I favor markets where steps in the supply chain are decoupled in such a way that competition at a given step will drive pricing down, lowering the total cost of the chain. I might rent aUHaul van, but I might avoid it if I were locked into buying gas only from a UHaul gas station along the route. This is one important aspect of the tension between modularity and interdependence. With this in mind, a foreclosure attack will work best when:
The combined steps make it easier to navigate a job "supply chain"
The total cost of the supply chain is lowered
I do not feel that I am placing future costs along the supply chain at risk
Apple's iTunes provides a great example. By rectifying the multiple steps between downloading, organizing, and synchronizing music, Apple made things far easier for consumers with the job of "I want an easy way to consume the music that interests me." Consumers were willing to pay a premium for this access, roughly a dollar per song more than could be obtained through music sharing sites. At the same time, Apples sales have been stunted by the layers of digital right management technology packed around each song file, repelling users (often power users) who worry that future supply chain costs are being placed at risk and who worry in particular than the value of a present purchase may erode if Apple changes its policies regarding music transfer to CD or some other media. This is a case of a great jobs to be done fit running into shortfalls on restrictions held by a specific user group. This is not normally interesting toInnosight, which avoids power users, but it is relevant to this paper.
So this paper is useful, because it provides a new avenue toward looking at competitive strategies and this new avenue can be (roughly)reframed in Innosight language even if the focus on power users makes this less relevant to Innosight core IP . That said, I'd like to recast this paper in the future in terms of new market and overshot users, both of whom work with product component valuations that sit below, rather than above marginal cost.
[(*)As a complete side note, tests for interaction demand large sample sizes. Online services have an advantage when sandboxing new products because it is easy to obtain large samples in a low cost way. This is not true for every product.]
Update: well, I was kinda wrong
Update 2: Something closer to a video e-card offering
===================================================================================
The Immigrant Technologist points to an interesting working paper written by several researchers within the HBS Entrepreneurial Management unit. The paper, "Platform Envelopment" takes a look at competition against entrenched platforms with large networks- from digital music players to online recruiters and shipping providers- where competitors may either compliment each other or share some functionality.
The researchers use a variety of case examples to generate three competitive approaches as follows:
A "conglomeration attack" where a company bundles together platforms that share a common user base but provide different services to those users. Examples includeDoCoMo offering payment platforms to mobile users who presumably have credit card accounts and phone companies offering bundled phone, cable, andinternet services
An "intermodal attack" where two companies offering similar (but not identical) platforms begin offering similar services. Examples include DHL offering overnight services in response to new FedEx offerings and Monster.com offering social network-style profiles as Linkedin launched a jobs board
A "foreclosure attack" where a firm begins offering a substitute service in order to place pressure on a complementary service offered by a competitor. Examples include Microsoft offering a "save as .pdf" option with Office 2007 in order to put pressure on Adobe and EBay offering Billpoint (until it failed) to put pressure on paypal.
The writers go on to ask about the factors that may impact success within each of these competitive approaches in the context of user valuations, asking whether a user will see more value in using two or more goods/ services together rather than separately. They suggest that this is a matter of relative valuations suggesting that two platforms should be bundled (conglomeration attack) together when the user valuations of one or both goods and services is substantially higher than the marginal cost of said goods and services at the profit maximizing price.
This suggests to me that the strategies are oriented toward a base of power users, if only because the "average user" should have a valuation closely linked with the price point at which the marginal cost and marginal revenue curves meet (ie, the profit maximizing price). This further suggests something that should be somewhat obvious: convergent services succeed when the user valuation of the sum of the products exceeds the summed valuation of each of the discrete parts. In statistics, we might identify these cases via a test for interaction (*). The paper (because it is a working paper) does not verify this assertion by defining the users (from power users to casual users) who will see the value in bundling.
"Aha!" I say. Consumer valuation is closely related to the "fit" between a particular solution and a job to be done. What if I were to re-frame this in terms of jobs to be done?
Conglomeration Attack: Here we have two functionally independent technologies. Users have adopted one technology to address one set of jobs and happily use a second platform to address a different set of jobs. In theDoCoMo case, I use a mobile phone to keep in contact with other people and I use a credit card to make purchases. In order for bundling to work in this case, the bundled platform will need to satisfy a new, previously unsatisfied job that is important to some consumer subset. It will not be enough to simply treat a phone as an extra credit card. My credit card jobs are suitably satisfied. It should not surprise me that bundling efforts often fail until they move away from consumers who already use both platforms and toward consumers who have access to one platform but not both. Phones become a channel formicropayments in cash based economies with low credit card penetration. To place this in the context of valuation and marginal cost, the MC/V of the bundle increases over the MC/V for each of the constituents because the dominant platform in the bundle (here, the phone) sharply lowers the marginal cost of access to the new component (purchasing) so that the value of being able to make a purchase without cash can suddenly exceed the cost of obtaining access to small amounts of credit.
Intermodal Attack: Power users are likely to use both competing systems concurrently, so switching costs will be low. A DHL power user will also have an online account with FedEx. The most demanding Monster.com users will also have LinkedIn accounts. These users will typically employ both services to address a focal job, whether "I want to reduce my worry when delivering critical packages" or "I want to be sure that I am not missing out on a new and interesting career opportunity." The services are left to compete in their ability to address these core jobs (actually, the jobs/ restrictiontradeoffs). It makes sense, then, that intermodal attacks often result in mirrored behavior. An improved strategy in this arena might result in designing new servcies that really fit consumer jobs to be done instead of just mirroring competing services.
That said, the LinkedIn jobs posting could really be recast as a conglomeration attack while the Monster social network could really be recast as a foreclosure attack. All of these approaches exist in a spectrum. UnlikeDHL/ Fedex, which are conceptually "closer," LinkedIn & Monster approach career networks from very different angles and this creates an asymmetry of switching costs. I imagine that it is more of a pain for aLinkedIn user to recreate a social network within Monster than it is for a Monster user to repost information in a larval profile on LinkedIn. At the same time, it is easier for a LinkedIn user to employ the LinkedIn job search service even while using the Monster.com service in parallel. This creates a market that is advantageous to LinkedIn in the long run even if LinkedIn faces real challenges in replicating Monster.com's voluminous jobs database and even if it is a pain for employers to repost inquiries on LinkedIn. In this case, the jobs are similar but the "ease of adoption" restrictions drive meaningful differences between the services.
Foreclosure Attack: Consumers often employ multiple goods and services in multistep processes in order to address a given job. If I want to get a large couch back from a retail store, I might employ a rental service, a van, a gas station and a dolly. I am often aware of the total cost of this "supply chain" and I favor markets where steps in the supply chain are decoupled in such a way that competition at a given step will drive pricing down, lowering the total cost of the chain. I might rent aUHaul van, but I might avoid it if I were locked into buying gas only from a UHaul gas station along the route. This is one important aspect of the tension between modularity and interdependence. With this in mind, a foreclosure attack will work best when:
The combined steps make it easier to navigate a job "supply chain"
The total cost of the supply chain is lowered
I do not feel that I am placing future costs along the supply chain at risk
Apple's iTunes provides a great example. By rectifying the multiple steps between downloading, organizing, and synchronizing music, Apple made things far easier for consumers with the job of "I want an easy way to consume the music that interests me." Consumers were willing to pay a premium for this access, roughly a dollar per song more than could be obtained through music sharing sites. At the same time, Apples sales have been stunted by the layers of digital right management technology packed around each song file, repelling users (often power users) who worry that future supply chain costs are being placed at risk and who worry in particular than the value of a present purchase may erode if Apple changes its policies regarding music transfer to CD or some other media. This is a case of a great jobs to be done fit running into shortfalls on restrictions held by a specific user group. This is not normally interesting toInnosight, which avoids power users, but it is relevant to this paper.
So this paper is useful, because it provides a new avenue toward looking at competitive strategies and this new avenue can be (roughly)reframed in Innosight language even if the focus on power users makes this less relevant to Innosight core IP . That said, I'd like to recast this paper in the future in terms of new market and overshot users, both of whom work with product component valuations that sit below, rather than above marginal cost.
[(*)As a complete side note, tests for interaction demand large sample sizes. Online services have an advantage when sandboxing new products because it is easy to obtain large samples in a low cost way. This is not true for every product.]
