Full circle to Sitecore – part 1

Part 1: Abandon ship! Goodbye monolith, hello headless

In this series of blogs, Sitecore veteran Bruce Davis-Goff talks about his journey from disillusionment with all-in-one monoliths, his leap to headless CMS and arriving full circle – back to Sitecore.

My journey with Sitecore started around 18 years ago with a new role as Web Manager for a prominent real estate company in New Zealand. Tasked with choosing an enterprise CMS, I narrowed it down to SharePoint (gulp), and a new player in the market, Sitecore.  Sitecore was the brainchild of a group of Danish tech-heads and was then, at version 6.3.

Legend has it, (this may or not be 100% factual), that a Danish Microsoft-centric company called Pentia A/S, was tasked with building a form-based XML editor. The story goes, someone got right carried away and ended up building a fully blown CMS prototype that induced a slight financial panic due to the hours involved. Sensibly, the company convened an all-hands with wives and partners and decided to go to market with the software, and the rest, as the saying goes is history.

Now one thing I have maintained consistently, is that Sitecore was in many ways, the original headless CMS. The main idea was that content, (stored as XML) was entirely separated from the presentation layer,  which at this point was Webforms (boo) and XSLT (yay).

This was at a time when every other CMS used proprietary methods to store content mixed with HTML. This meant the content was only usable by the CMS itself and not any external systems. Sitecore, on the other hand stored the content as pure XML, you could do what you wanted with it. While retrieving the content was nowhere near as easy as today’s Json based microservices, it was a big leap forward for content abstraction and reuse.

I will confess right here and now that I was a big fan of XSLT; the language and examples were well documented, it was easy to use and deploy, (no compiling) and changes could be rapidly tested. Plus, I’d never written in a compiled language such as C#, so it really was the path of least resistance, or to make a bad coding joke, for-each to their own.

The pace of development on the Sitecore platform was truly amazing, new features and releases arrived quickly. With this, a strong user community formed and it’s fair to say that a lot of advancements and enhancements have come from the talented developer and user group community, many who ended up working for the mothership.

It wasn’t long before being a CMS (admittedly brilliant) was nowhere near enough, and so the marketing features started appearing, personalisation, A/B Testing, Path Analysis, Dynamic Segmentation, and much more. All of this depended on the ability to capture and store near-real time analytics at scale. MSSQL didn’t really have the chops for this back then, and maybe this is where the sprawling beast that Sitecore XP became really started to emerge with the addition of Mongo databases to handle analytics – there was a bit of changing horse mid-stream that went on for sure.

Today’s Sitecore XP is big – really big and has so many tricks up it’s sleeve you could spend, oh I don’t know, 18 years working on the platform and still be learning new stuff.  With that big footprint, comes big infrastructure requirements, the whole stack is Microsoft based so the hosting costs can also be truly eye-watering.

It’s here that something of a disconnect happens for a lot of Sitecore customers who have spent a small pacific island’s budget on a thoroughly comprehensive product. It’s a few years later and  they are only using 10% of the functionality. Sort of like buying a fully stocked chef’s kitchen when you just need to boil an egg.

The other point I often make, is a large monolithic system can be likened to a one-man band,  jack of all trades, master of none – though I’m happy to state I think Sitecore does master data modelling and content management. However, it’s a big ask to make one product that rules them all, (just ask Sauron), and agile competitors can bust out “features-to-market” quicker than  Sitecore XP any day.

And it’s at this point that the monolith makes no sense, even to enterprise customers. In the headless world of instantly available, high-speed micro-services – why buy the kitchen, when you just need a stove, a pot, and some water? If you need anything fancier, you can pick and choose from the best products on the shelf and not have to buy the tripe which you know you will never cook.

At this point, Sitecore’s SaaS offering, XM Cloud not yet ready, and it was here in my journey that I threw my life raft overboard and abandoned ship.  After 15 years, of working with Sitecore, working for Sitecore as the NZ business development manager, training Sitecore developers, running and funding the NZ user groups, developing and blogging and being part of the amazing community – it was over. That ship wasn’t turning anywhere near quickly enough and I was really questioning why I had given so much to a company that gave back so little.

But that’s a tale for another day, my little life raft washed up on the shores of a beautiful tropical island called Konabos, the outstanding Sitecore specialist company headed by veteran gurus and all-round great guys Akshay Sura and Kamruz Jaman. They too had been taking side missions and together we ploughed headlong in the new headless world.

First stop was finding what an alternative to Sitecore might look like and understanding how content modelling worked in the new world. The inheritance-based modelling of Sitecore was something I initially missed, where the blueprint for content item can be made of smaller, simpler templates, allowing great flexibility and eliminating field duplication.

There are a million little players in the headless space and more than a few giants. Professionally, after much evaluation our pick was Kentico Kontent AI for decently size projects,  Contentful for everything else. Choosing React / Next.js as the language / framework of choice for the foreseeable future came out in the wash as well.

The last exciting piece of the puzzle was Vercel and Netlify, the ability to plug a headless CMS and a Github repo into a CDN  and have multiple environments running in a matter of hours was a true breath of fresh air.  Plug in an enterprise predictive search such as Algolia, a world class DAM such as Cloudinary and you could have a slick web presence for the price of a good take-out coffee each month.   Did I mention blindingly fast? 90+ Lighthouse scores, bombproof security and 100% uptime? These things are only achievable on an Azure stack with much effort and cost.

Exciting times indeed, I had moved on and Sitecore was looking like a pig with lipstick. I was having fast times with my new love and I wasn’t looking back.

“But”, I hear you saying, “surely that’s not the end of the Sitecore story?”. And you’d be right – for a good while I though I’d be spending my remaining days offering Sitecore palliative care,  but there’s always mouths to feed, and the twists and turns of life are manifold.

Like Hotel California, or an organised crime family, just when I thought I was out, they dragged me back in. How did that go?  Join me in the next episode where I pack up my pencil case and head back to school for Sitecore Training Revisited.