Blog post: Platform Engineering Through the Lens of Two-Sided Markets #515
Conversation
nagyv
requested review from
AloisReitbauer,
AlexsJones,
thschue,
joshgav,
lianmakesthings,
kgamanji,
justincormack and
angellk
as code owners
✅ Deploy Preview for tag-app-delivery ready!
To edit notification comments on pull requests, go to your Netlify site configuration. |
|
Great, thanks for opening this @nagyv ! I think you'll need to manage the DCO failure, there are some instructions for how to do that in the CI step. In addition, it would be great to remove the "draft" status so we can see it on the blog preview (to review headings etc). @joshgav / @roberthstrand it would be great to get a review of this! |
Signed-off-by: Viktor Nagy <viktor.nagy@gmail.com>
"Improve the style of the following article. The article is to be published on the Cloud Native Foundations App Delivery TAG's blog. Article:" Signed-off-by: Viktor Nagy <viktor.nagy@gmail.com>
Signed-off-by: Viktor Nagy <viktor.nagy@gmail.com>
Kept only disclaimer and TL;DR Signed-off-by: Viktor Nagy <viktor.nagy@gmail.com>
nagyv
force-pushed
the
nagyv-two-sided-markets-blog
branch
from
5cb14be
to
1598003
Compare
website/content/en/blog/the-economics-of-the-cncf-platform-engineering-model.md
Outdated
Show resolved
Hide resolved
website/content/en/blog/the-economics-of-the-cncf-platform-engineering-model.md
Outdated
Show resolved
Hide resolved
| I called these platforms many sided. The [CNCF Platform White Paper](https://tag-app-delivery.cncf.io/whitepapers/platforms/#why-platforms) describes the motivations why enterprises might invest into platforms: | ||
|
|
||
| “The desire to refocus delivery teams on their core focus and reduce duplication of effort across the organisation has motivated enterprises to implement platforms for cloud-native computing. By investing in platforms, enterprises can: | ||
|
|
There was a problem hiding this comment.
I'm struggling to understand how this is a two-sided market problem. A platform team can be expressing opinions, like focusing on some golden paths, that are the recommended approach to specific problems. This seems more like a highly curated store than a true two-sided market to me.
There was a problem hiding this comment.
Two-sided markets are not defined by their offering, but by the market dynamics.
Engineers do their work with or without a platform. Why would they switch to a platform? Because it offers them value. The more value it offers, the more engineers switch. The more engineers switch, the better the platform can become. The more features the platform supports, the harder it is to replace it.
This is the two-sided market dynamics. As you increase usage on one side of the market, there is more reason to increase usage (in this case support) on the other side of the market).
There was a problem hiding this comment.
I wonder if the confusion is that the examples provided are focused on actually higher value one-sided market production but being presented as two-sided?
What I mean by that is that the examples in the article as well as in this thread speak about making a higher value offering by the platform teams to the user application teams. This would be a one-sided offering in that the platform is offering a solution to the users and not that the platform is offering a solution that focuses on connecting producers and consumers.
To take this away from platform engineering for a minute, an example of a platform that is "one-sided" may be Netflix. It offers users movies, but it does not specifically help content creators distribute via Netflix without the explicit purchase/involvement of Netflix in that process. In contrast, Etsy would be a two sided market. The value of Etsy is in how it enables buyers and sellers to meet. Buyers are provided tools by the platform to advertise, communicate, and distribute their wares while consumers are provided tools to browse, investigate, and purchase their needs. Etsy does not improve by creating better ceramics, it is improves by creating better tooling for one or both sides of the market.
I think the quote this thread speaks to talks about the desire for the enterprise to provide a solution to users that meets their needs. It does not specifically speak to if the enterprise does this in a one-sided or two/multi-sided solution.
So in conclusion, this comment is 100% true of two-sided markets, but also of any product with consumers
This is the two-sided market dynamics. As you increase usage on one side of the market, there is more reason to increase usage (in this case support) on the other side of the market).
And it seems in the case of this example, the fact that support would be increased, it is maybe pointing more towards a one-sided market since it is the producers of the platform that are increasing investment, not a different platform consumer/producer doing so.
There was a problem hiding this comment.
@abangser I like your Netlifx - Etsy examples a lot, and I agree with everything in your comment except for the last half-sentence.
the fact that support would be increased, it is maybe pointing more towards a one-sided market since it is the producers of the platform that are increasing investment, not a different platform consumer/producer doing so.
This will really depend on the organisation structure. Who owns the platform (e.g. Backstage installation) vs Who owns the scaffolding templates and integrations? If these are owned by the same team, I agree with you. If these are owned by different teams, then the more users use the platform as consumers, the more compelling it is for "supplier" teams to join the platform. Support of the platform itself does not increase, the platform owners job does not change.
|
|
||
| ## Adopting a platform | ||
|
|
||
| As already noted, a central question of multi-sided markets is their adoption. The more users are on a platform, the more investment goes into a platform. The more features a platform offers, the more users want to use a platform. But how does one start this cycle? Where is the turning point that will start a self enforcing spiral of adoption? |
There was a problem hiding this comment.
This sounds more to me like product-market fit, common to all products. Again, I'm not grokking why this is specific to a 2-sided market.
There was a problem hiding this comment.
@earth2marsh Product-market fit is about feature adoption, not platform adoption. AFAIK.
There was a problem hiding this comment.
IMHO the key is who is doing providing the "more features a platform offers".
If it is the creators of the platform, then I would agree with the point @earth2marsh has said in that it is a genearl product fit/investment. If it is producers who are enabled to provide features via platform tooling, then I would agree with @nagyv that this is a 2-sided market.
There was a problem hiding this comment.
@earth2marsh @abangser I might be able to better explain this through Customer Lifetime Value. Let's assume that our "platform" has a market of fully identical users. The first user joining bring CLV €100.
On a "regular" market, the 2nd user joining will have a CLV €100 as well, and the CLV of the first user does not change.
On a networked market (and every two-sided market has network effect, but not every networked market is two-sided), the CLV of the first user has two components:
- their own CLV
- the value their presence brings to the following users
Thus the 2nd users will attribute higher than €100 to the platform, even though their CLV is €100 (+ the value their presence brings to the 3rd+ users)
In the Internal Dev Platform area, as argued in this article, there are multiple sides to the platform. The platform in itself is just a skeleton, a service registry, a scaffolding engine.
- a platform team starts to develop templated for the platform
- this brings value to developer users
- some developers join the platform
- this brings value to the already present platform team AND
- this brings value - for example - to the security team
- the security platform team joins the platform to add their own tools
- ...
Does this help to understand the two-sided nature better?
|
Thank you for this @nagyv! I did an initial review now and will read again soon, a couple thoughts: Is there an opportunity to discuss "Thinnest Viable Platform" in the context of build versus buy? Could we emphasize more that even if an organization buys most of a platform from GitLab or another vendor they'll still be responsible for the "thinnest layer" on top, even if that's just documentation pointing to e.g. GitLab? I'm also a bit concerned that the only vendor mentioned is GitLab and @nagyv you're a PM for GitLab 🤣 - perhaps it would come off better to mention another vendor or two as well - WDYT? Thanks again for the submission! |
joshgav
added
wg-platforms
|
The title of the post is currently |
There was a problem hiding this comment.
I believe the other reviews in this PR have highlighted two possibly underpinning challenges to publication:
- The title is about two sided markets, but we drift quite a lot from that topic which shows both in the file name (focuses on economics) and the initial summery paragraph (focus on getting started, value loops, and cost of change).
- This article comes with the disclaimer of a vendor author and will contain a marker that it is a community submission, but it does lean a bit far towards vendor pitch.
IMHO neither of these need to be blockers to publication. In particular, I added a number of small suggestions in this review that would soften the vendor bias and I think make that easier to publish under the CNCF.
I think the harder one is the inconsistencies and possibly disagreements over the two sided market. The really good questions from @earth2marsh led me to return to an article I read on 2-sided markets just recently and realise that this article fundamentally describes two-sided markets differently. I have posted quite a long comment on the subject that may help clarify my concern so maybe we can pick up the conversation there.
I do want to reiterate, this will be posted as a community submission and therefore is open for a follow up from others if anyone feels a clarification could be in order. So while I think it is important we bottom out that conversation within reason, it is also important that we leave space for opinions in this community blog post format.
|
|
||
| To summarise the economics: In two or multi-sided markets, the products are platforms with users and suppliers who evaluate the platform: the more or higher quality users make a platform more compelling to enter for the suppliers; the more or higher quality suppliers make the platform more compelling to enter for the users. | ||
|
|
||
| For the rest of the article, I’ll speak about multi-sided markets as they better capture the world of DevOps platforms, and the economics in our case is very similar to two-sided markets. |
There was a problem hiding this comment.
This seems contradictory, you say "I’ll speak about multi-sided markets as they better capture the world of DevOps platforms" but then immediately follow up with "the economics in our case is very similar to two-sided markets".
So do you mean to say two-sided and multi-sided markets are the same? Or is one more of the focus for this article? 🤔
website/content/en/blog/the-economics-of-the-cncf-platform-engineering-model.md
Outdated
Show resolved
Hide resolved
| I called these platforms many sided. The [CNCF Platform White Paper](https://tag-app-delivery.cncf.io/whitepapers/platforms/#why-platforms) describes the motivations why enterprises might invest into platforms: | ||
|
|
||
| “The desire to refocus delivery teams on their core focus and reduce duplication of effort across the organisation has motivated enterprises to implement platforms for cloud-native computing. By investing in platforms, enterprises can: | ||
|
|
There was a problem hiding this comment.
I wonder if the confusion is that the examples provided are focused on actually higher value one-sided market production but being presented as two-sided?
What I mean by that is that the examples in the article as well as in this thread speak about making a higher value offering by the platform teams to the user application teams. This would be a one-sided offering in that the platform is offering a solution to the users and not that the platform is offering a solution that focuses on connecting producers and consumers.
To take this away from platform engineering for a minute, an example of a platform that is "one-sided" may be Netflix. It offers users movies, but it does not specifically help content creators distribute via Netflix without the explicit purchase/involvement of Netflix in that process. In contrast, Etsy would be a two sided market. The value of Etsy is in how it enables buyers and sellers to meet. Buyers are provided tools by the platform to advertise, communicate, and distribute their wares while consumers are provided tools to browse, investigate, and purchase their needs. Etsy does not improve by creating better ceramics, it is improves by creating better tooling for one or both sides of the market.
I think the quote this thread speaks to talks about the desire for the enterprise to provide a solution to users that meets their needs. It does not specifically speak to if the enterprise does this in a one-sided or two/multi-sided solution.
So in conclusion, this comment is 100% true of two-sided markets, but also of any product with consumers
This is the two-sided market dynamics. As you increase usage on one side of the market, there is more reason to increase usage (in this case support) on the other side of the market).
And it seems in the case of this example, the fact that support would be increased, it is maybe pointing more towards a one-sided market since it is the producers of the platform that are increasing investment, not a different platform consumer/producer doing so.
|
|
||
| ## Adopting a platform | ||
|
|
||
| As already noted, a central question of multi-sided markets is their adoption. The more users are on a platform, the more investment goes into a platform. The more features a platform offers, the more users want to use a platform. But how does one start this cycle? Where is the turning point that will start a self enforcing spiral of adoption? |
There was a problem hiding this comment.
IMHO the key is who is doing providing the "more features a platform offers".
If it is the creators of the platform, then I would agree with the point @earth2marsh has said in that it is a genearl product fit/investment. If it is producers who are enabled to provide features via platform tooling, then I would agree with @nagyv that this is a 2-sided market.
website/content/en/blog/the-economics-of-the-cncf-platform-engineering-model.md
Outdated
Show resolved
Hide resolved
website/content/en/blog/the-economics-of-the-cncf-platform-engineering-model.md
Outdated
Show resolved
Hide resolved
website/content/en/blog/the-economics-of-the-cncf-platform-engineering-model.md
Outdated
Show resolved
Hide resolved
website/content/en/blog/the-economics-of-the-cncf-platform-engineering-model.md
Outdated
Show resolved
Hide resolved
website/content/en/blog/the-economics-of-the-cncf-platform-engineering-model.md
Outdated
Show resolved
Hide resolved
…ineering-model.md Co-authored-by: abangser <bangser.a@gmail.com> Signed-off-by: Viktor Nagy <126671+nagyv@users.noreply.github.com>
…ineering-model.md Co-authored-by: abangser <bangser.a@gmail.com> Signed-off-by: Viktor Nagy <126671+nagyv@users.noreply.github.com>
…ineering-model.md Co-authored-by: abangser <bangser.a@gmail.com> Signed-off-by: Viktor Nagy <126671+nagyv@users.noreply.github.com>
There was a problem hiding this comment.
Really enjoyed following the discussion on this PR!
Seeing as this will be a public blog post, I wonder if there visual elements like diagrams, charts, or graphs to illustrate the concepts. For example, a visual representation of the CNCF Platform Engineering Maturity Model could enhance understanding. This could also encourage reader engagement by asking questions or prompting them to share their experiences related to the topics discussed.
You referenced several external sources, such as the CNCF Platform White Paper, the CNCF Platform Engineering Maturity Model, and specific tools like Auto DevOps from GitLab and Backstage from Spotify. Would it be helpful for readers to have easy access to links citated?
Just my 2 cents --- thanks for the write up!
There was a problem hiding this comment.
Just wanted to confirm I think that this meets current blog expectations.
I believe it is a bit close to a vendor view, but I am very grateful to @nagyv for his responsiveness to both questions, ideas, and suggestions.
Thank you for bringing this idea to the group!
|
Hey @nagyv, happy new year! I know the end of the year ended up a bit hectic and bogged down by illness. I am hoping you are still interested in posting this blog and I would love to continue supporting you if there is anything I can do? For example, we definitely still need a final review from TAG leadership (I think @joshgav had this on a todo list somewhere?). |
There was a problem hiding this comment.
I have some small, mainly stylistic suggestions.
@abangser and @earth2marsh just making sure: Have all your concerns been sufficiently addressed?
|
|
||
| ## The theory of platforms | ||
|
|
||
| In economics platforms are modelled as [two or multi-sided markets](https://en.wikipedia.org/wiki/Two-sided_market). Basic economics speaks about markets as the interaction of suppliers and consumers. A two-sided market might still have suppliers and consumers, but the market itself is a product with its own market. A few well-known two-sided market products are operating systems or social networks. There is a market for operating systems that consumers choose. At the same time there is a market for operating systems that application developers choose. The more applications are available for a given operating system, the more users want to choose it; and the more users use an operating system, the more applications are developed for it. Interestingly, this aspect of two-sided products brings interesting market dynamics with it. |
There was a problem hiding this comment.
Interestingly, this aspect of two-sided products brings interesting market dynamics with it.
Would drop the first word here to avoid repetition
website/content/en/blog/the-economics-of-the-cncf-platform-engineering-model.md
Outdated
Show resolved
Hide resolved
|
|
||
| While physical markets might have similar characteristics, because of the nature of the physical world (travel is costly) the two-sided market considerations are often negligible. On the other hand, for digital goods a two-sided market suffers for several properties that need regulatory intervention. | ||
|
|
||
| To summarise the economics: In two or multi-sided markets, the products are platforms with users and suppliers who evaluate the platform: the more or higher quality users make a platform more compelling to enter for the suppliers; the more or higher quality suppliers make the platform more compelling to enter for the users. |
There was a problem hiding this comment.
the more or higher quality users make a platform more compelling to enter for the suppliers; the more or higher quality suppliers make the platform more compelling to enter for the users.
Again, I find this sentence hard to follow, stylistically. Also what exactly do you mean by "higher quality"?
There was a problem hiding this comment.
@lianmakesthings This section is part of the introduction to two-sided markets, as a result "higher quality" is market dependant. e.g. on Facebook, higher quality on the advertiser side means more relevant, less intrusive ads; higher quality on the Facebook "user" side likely means more buying power.
If you have suggestions for alternative wording, I'd be interested to see it.
|
Thanks for the review @lianmakesthings. I think that blockers have been met by the disclaimer, additional tool examples, and softening of language around defining the category. I would love to see @nagyv get this published. With that in mind, IMO all open suggestions (including yours @lianmakesthings) should be optional as at the end of the day this is a community post and while different people have opinions this post appears to meet community guidelines and if a community member wants to write on a similar topic with a different voice or from a different angle that would be also be welcome. I know it has been a long road @nagyv and that the start of the year can be a hectic time but please do let me know what I can do to help get this published for you! |
|
@nagyv it would be great to see this merged - could you review @lianmakesthings 's comments please? |
|
Hey @nagyv, any word on if/when you may want to review this for merge? Would love to get this on the site! |
|
Sorry everyone, I got distracted by life. I'm back on this now, and hope to have everything answered this week. |
Signed-off-by: Viktor Nagy <126671+nagyv@users.noreply.github.com>
Fixed |
Closes #501