Last time around in my 2020 predictions, I did reasonably well, so I’m going to be a little bolder in my 2021 set. To be clear, I think some of these may be a little too early (i.e. ad network consent as a feature, SPAC exit, etc.) but here goes.
Previous entries: Part 1 (2020 prediction review) and Part 2 (what I missed in 2020).
The WCM/DXP market will see its first exit via SPAC
A lot of PE money has gone into WCM/DXP vendors in the last few years (Sitecore, Episerver, Acquia) and in the past, there were three exit options;
- Sell to a larger player in the market. This has been the model in the past – for example, Vista flipping Marketo to Adobe after a year of tightening the operations there (for a massive return on such a short time frame).
- Sell to another PE firm. TCV flipped Sitecore to EQT. Accel-KKR flipped Episerver to Insight Venture Partners. The problem is; each new PE firm in the chain has to believe that they can continue to optimize or grow that business, and the price tag gets larger each time. I suspect that we’re nearing the end of that chain for a few vendors.
- Go IPO. This has always been the hardest option – especially in a space where the growth for the overall market is 16%. The only companies that have managed to do so (or likely will) are those with a strong B2B brand or consumer-facing motion, such as Wix or HubSpot. Squarespace or Netlify may be able to take advantage of the current “narrative bump” due to Covid (every IPO needs a solid narrative – check out this podcast from Barry Ritholz and Nobel Prize winner Robert Shiller on the topic) and potentially go down this route.
However, there is a newer financial package model of late, and that is the SPAC. In short, at the moment, there are over 200 SPACs, armed with over $70 billion in funding, and in general, that money has to find a home – and fast (i.e. within two years; no SPAC wants to have to “give up” and simply return the money to shareholders; with two years of negative return against inflation).
In many ways, the SPAC is the ideal route for some of that PE and VC money – there is far less of a “narrative hurdle” compared to an IPO and the current climate means that solid, but unexciting businesses will likely seriously consider this route over any other. Indeed Thoma Bravo (which has some investments in the space via Hyland and others) has already announced a $900 million IPO for a SPAC they are launching.
SaaS vendors may be slow to move to the Enterprise DXP market, but the Enterprise DXP market will move to “SaaS”
In my original post of 2020 predictions, I noted that SaaS vendors will be slow to take on the enterprise DXP market. One of the major reasons for this is that in order to be effective with a SaaS WCM product, you either need to accept the limitations in the product offering, or be able to build significantly on top of an API-first platform (usually with a great deal of in-house resources). Because of these two limiting factors, many buyers looking to address large digital projects still look to solutions such as Sitecore, Acquia, Episerver and Adobe – which provide a lot of product functionality out of the box.
Each of those vendors have made different product and architectural moves of late;
- Adobe has done significant work re-architecting AEM around a more scalable PaaS offering (AEM as Cloud Service) – but the fact that Adobe has partnered with IBM to allow them to host the “old” AEM offering on their cloud seems to indicate the uptake has not been that great.
- Sitecore abandoned their “Sitecore SaaS” transformation, and are now looking to build more SaaS features on top of the Content Hub acquisition (aka Sitecore Edge). Similarly, they also removed PaaS support entirely in the upcoming XP 10.2 release in favour of better container support.
- Episerver acquired Optimizely, which should bring additional SaaS capability and expertise into the organization.
- Elastic Path acquired Moltin, for many of the same SaaS-centric reasons.
Full multi-tenancy SaaS theoretically has no impact on the end customer (but some will pay for single tenancy specifically to ensure consistent performance, or for legal/security reasons) – in practice, if a vendor were running a really well-done managed service with many of the friction pain points removed, it may be more than sufficient for many. However, multi-tenancy also theoretically means that a software vendor can run with lower infrastructure costs, meaning they can either buy market share through reduced costs on their customers, or redirect that additional savings into R&D or other parts of the business.
In practice, it may mean that vendors start to run hybrid approaches; where elements of the offering are full multi-tenant SaaS; for things like the content repository, personalization/optimization, image transformations, etc. while providing a tightly integrated means of building the implementation on top.
Some DXP vendors are already moving this way (largely by acquisition – see Episerver/Optimizely or Sitecore/Stylelabs); headless vendors need to be able to start to offer a more complete package for the entire infrastructure (i.e. a place to run Contentful apps, SSG, deployment, etc.). Preston So describes this as a bifurcation, but I see most (smart) vendors looking to reconcile those marketer and developer needs by addressing both with architectures that make sense for the task.
Jamstack starts becoming less interesting than the stuff built on top of it.
One of the advantages of being around so long (like Bill Murray in Groundhog Day) is that you can start to see parallels in technology trends.
I am particularly interested in the idea of Jamstack and how this is related to the LAMP trend. LAMP was an architectural pattern which stood for Linux, Apache, MySQL and PHP and it became a pattern for many of the early complex web applications. LAMP was fairly revolutionary and unlocked a great deal of value by doing a number of things;
- Creating well understood patterns between architectural tiers; you had your OS, web server, database and programming language.
- Well maintained APIs between those layers – you had out of the box APIs that would allow each layer to properly abstract and have known interfaces on the other components in the stack.
- This, in turn, allowed developers to easily scale quickly to build applications on top of this well-understood pattern.
Enter Jamstack. It stands for JavaScript, APIs and Markup. This was a term coined by Matt Biilmann, the CEO of Netlify. However, it’s outgrown the name somewhat, since, in practice it generally refers to the following;
- JavaScript (language) either on the front-end, or via static-site generation (i.e. Next.JS, Gatsby)
- CMS (content database – i.e. Contentful, Contentstack, Kentico, Sanity, etc. or something Git or file based)
- Hosting service (infrastructure – i.e. Vercel, Netlify, Heroku )
Which can compare somewhat to how LAMP was understood:
- PHP (language)
- MySQL (database)
- Apache/Linux (infrastructure)
As the patterns for how these interact become more commoditized and understood, ultimately, the architectural elements of the framework will become less relevant than the software that will be built on this framework. Currently Jamstack is the story, as much as “LAMP” was the story in the early days. But over time, the focus was far less on what that particular architecture enabled, and the types of applications built on top of it. From the LAMP stack, the world got content management software like Drupal, WordPress and Joomla. Today, nobody talks about that underlying stack, they simply talk about the application and what needs it is fulfilling, and the development task for many is often replaced by scripted services and the task of installing plug-ins and themes.
Jamstack will probably get to a similar point. What this means is that current headless vendors will have to enable meeting complex application needs via the community and platform (in the same way that WordPress and Drupal did), or other tools will start to fulfill specific use cases on top of Jamstack principles. This leads to my next prediction…
Low-code/No-code and other enabling frameworks become more important
As I note in my previous prediction, there are more and more vendors like Frontastic, Mobify, Uniform and Nacelle that exist entirely on top of other technologies. Other vendors like Shogun, Netlify and Cloudflare are taking the approach of optionally being able to replace some parts of the stack, such as CMS, while still having a specific other business problem they are solving.
I also fully expect where other vendors have built specialized content repositories for things such as knowledge bases or chatbots, a new breed of vendor based on Jamstack principles starts enabling these use cases on top of other vendors; you get the best of both worlds; out-of-the-box, specialized expertise for a use case, built on a re-usable, cross-channel framework.
The need for this “best of both worlds” approach; namely a strong foundational API *and* being able to meet marketer and casual contributor needs is clear. The big question is whether existing vendors (both headless and legacy) start to enable this in their products, or if other new vendors start to fill those gaps for them. To be clear, there are still some major gaps to fill in the DXP space. Forrester has long talked about “Systems of Record” and “Systems of Engagement” but the fact is these need to play nicely with each other and often fail to add value (i.e. where insight helps inform a content or channel strategy). It is often left to very skilled practitioners (along with their tool of choice; spreadsheets) to make that work, and I see tools like these and other approaches start to make those connections.
Of course, I’m not the first to this trend – Scott Brinker has covered it for quite some time and I already see tools like Airtable starting to take over more “jobs to be done” within marketing organizations.
Ad-tracking becomes more transparent (for the betterment of both sides)*
Currently, Apple and Facebook are engaged in a virtual pissing contest, where Apple is positioning privacy as a consumer differentiator, and Facebook is trying to claim that small businesses will suffer due to being unable to target ads locally.
(While both are corporate behemoths, Apple has the slightly better position given that they are explicitly supporting the interests of their customers – i.e. the people that pay them money for products, while Facebook is making clearer the audience is the product and that businesses are their actual customer).
Regardless of this and other legislation undercutting the ad-revenue model, it seems that an obvious way forward is to make buyer interest and intent far more transparent and clear. As a customer, I am constantly bombarded by ads for things that I just bought; and this doesn’t work for me, or the advertiser. In fact, the only group it does work for is any ad network; they get paid to serve them up regardless of efficacy. Similarly, often the algorithms are never quite right. For example, my current boring car of choice is my Volvo station wagon (for the dog in the back). I’d really love for my next car to be electric (or optionally a plug-in hybrid) but it needs to be AWD just to get up my driveway. I am interested in any brand that meets this criteria; but instead, I see ads for Volvo SUVs (not interested) because advertisers are taking the brand view, rather than trying to understand or solve my actual problem.
Now, to be fair, this requires a detailed level of understanding my life situation that simply cannot accurately be determined by my browsing patterns across car sites and blogs. This may be more of a long-term bet (which is why I have the asterisk), but I would hope that a platform arises where consumers can actually clearly identify interests. In exchange, companies like Facebook don’t need to creepily follow me around. It’s clear that a lot of tech companies seem to have this socially stunted view of the world; instead of simply *asking* they default to creepier and more invasive forms of tracking in order to “replicate” a level of engagement. Yet they continue to get it wrong; poor segement identification (I continually get ads in French, solely because I live in Quebec; and once you are tagged with a trait, it follows you everywhere).
While I would prefer that companies start to treat this value exchange between consumers and advertisers more proactively, I suspect it will instead come out of necessity due to legislation, and ad-based businesses will fight it tooth and nail, but despite this, the end result will be better outcomes on both sides.
Amazon (and others) will start to further push into the DXP space (and what this means for headless vendors)
Amazon has already “productized” a few elements in their platform, including personalization (AWS Personalize) and headless content management (as a feature in AWS Amplify). I fully expect that would start to build on this to even further push into full-blown DXP and PIM scenarios.
Amazon probably sees something amusing about a company running a cloud PIM and DAM off Amazon Web Services in order to push products info to… Amazon, which is running on …Amazon Web Services.
In other words, why not cut out the middleman?
Jeff Bezos already owns a WCM system in ARC Publishing, so it would not surprise me to see them building out the capabilities to assist manufacturers and retailers selling on Amazon. In fact, as Amazon starts to actually compete with them in some cases (pushing down revenues), they need to sweeten the pie by pushing down costs, and this would be a major way to do so.
Of course, the risk to any API-first vendor is that they simply get made into a feature of a larger platform. The industry has a long history of this happening; Benedict Evans has a great post entitled: “Platforms, bundling and kill zones”
This is also happening by other adjacent vendors in the stack. Netlify has a CMS, Cloudflare has a CMS. Google, by virtue of controlling schema.org can also ensure similar adherence to standards to support their drive to dominating the SERP (Search Engine Results Page) experience – and as I have mentioned, it’s a short leap from schema.org to actual content modelling. In the past, Google has been content to merely influence the WCM vendors to try to support these larger efforts, but they do have a history of dipping their toe into pretty much everything. (Whether they actually support something like this long-term is another matter – i.e. this now irrelevance of Blogger, etc.).
Headless vendors have been a positive disruptive force in the industry. They have evolved to simultaneously meet the current multi-channel needs of most complex businesses, while also taking advantage of the major change in how front-end experiences are created (in a word, JavaScript frameworks). However, unless they evolve to start meeting more business scenarios (and having an opinion in what scenarios they can do better than anyone else) they will be reduced to being a fancy, commoditized, content databases.
Which brings me to my last prediction:
At least one headless vendor is acquired
There are plenty of vendors out there looking for SaaS expertise in CMS in order to round out portfolios; Salesforce CMS has been underinvested. SAP, Oracle and Microsoft, all need to have a stronger tie between content and marketing efforts. Adobe can afford to run two product tracks (as per my wildcard prediction) and larger platform players in services and infrastructure want some sticky product offerings. I could see the following vendors as potential buyers;
- Any cloud/hosting provider lacking that capability; GoDaddy, Rackspace, Pantheon, Salesforce Heroku to compete with Netlify, Cloudflare.
- Any suite provider looking to accelerate cloud CMS offerings in a larger martech stack (Salesforce, Adobe, SAP, Oracle, Microsoft)
- Any PE firm looking to do a rollup with other portfolio companies or enhance their current offering:
- Vista: Acquia, Gainsight
- Index Ventures: Gatsby, Strapi
- Insight Venture Partners, Episerver, Contentstack, Commercetools
- SaaS vendors focusing on the SMB and mid-market looking to move either upwards or more developer-focused (Shopify, Hubspot, Automattic?, etc.)
There are a number of headless vendors that are still quite small and not inflated by valuation rounds. I could easily see a larger vendor in need of some fresh ideas undertaking this as something of a “value-added acquihire”.