Hello world – the all new diginomica

SUMMARY:

diginomica has a fresh coat of paint and a whole lot more going on underneath the covers. Check out our case study on improving a media presence.

The HCM cornerstone page
The HCM cornerstone page

As I say on our video and podcast shows: “Hello boys and girls, sports fans…Den here” though this time I’m speaking on behalf of the whole diginomica team.

As mentioned at the back end of last year, we’ve just completed a drains up rebuild at diginomica. I dunno about you but the team is pleased with the outcome. What did we do and why?

Wassup doc?

After just shy of three years, diginomica was creaking. We say that in the nicest possible way because we could never have known back in the day just how well things would go and the impact that would have as we spent month after month patching the site, adding bits and pieces here and there. In the words of Workday’s Mark Nittler, diginomica became Frankensoft, in part because yours truly codes…badly. And while Den was banned from writing PHP and CSS codelines about a year or so ago, that didn’t stop him from mucking things up on occasion. What did we do?

Last fall, we spent a good amount of time finding the right people for the job because we knew from bitter experience that there are plenty of talkers and not so many doers. Our primary goal was to find a partner who would not only work with us but stick with us, providing ideas with which we could run. We found that partner in Devrix, who are based in Sofia, Bulgaria. Den knew they’d be good because he’d worked with Bulgarian coders before. They build damn fine code and that, for us, was critical.

More to the point, they came up with a design which we believe is both modern and sustainable. The bonus for us is that part of the Devrix team are core WordPress committers. In short, they know this stuff inside and out. Think in SAP terms. It’s like having a 15 year plus SAP dev on the team. Think in Oracle terms, it’s like having an Oracle Ace on the team. You get the picture.

Cornerstone pages – choose your topic and community, subscribe and prosper

One of our key objectives was to establish topic cornerstone pages. We believe that while many folk like a lot of our content, there are plenty who want to concentrate on one or more topic areas. That’s what the cornerstone pages deliver. Each is hand curated by one of the diginomica core team so that there is always something fresh to consume.

You can browse these 18 cornerstone topics via our three main topic menus:

We’re still spit-polishing a few of these areas for you as of this morning and ironing out nits, so if you see some weirdness then that’s something the devs are cleaning up. On each topic page you’ll see a chance to subscribe by email or RSS, and to browse the hand-curated editor’s picks.

If you’re coming here for the first time and would like a digest of specific content then feel free to click on one of the menu items to discover how we’re organizing the content. There’s an example at the top of this story.

More goodness

Another key objective was to improve search. We’ve done that using elastic, faceted search. It’s wicked fast. Check it out; it is a huge improvement.

We’ve introduced inline commenting. That means you can drop a comment adjacent to any paragraph. We think that makes commenting more meaningful and we hope readers will like that feature.

Yet another objective was to ‘cut the crap.’ Like many others, we’d added and added ‘stuff’ to the site but as anyone who has run open source software will tell you, there are no free lunches. Got free code? Great. Get it supported? Not so great. Pay for code? Great. Keep it supported? That’s a movable feast. Putting everything in our partner’s hands means we have one throat to choke when things go pear shaped, one partner that knows its code. We think that’s vital.

We also wanted to improve the mobile experience. It was already good but was not great. We think we’ve achieved that, paring back on all the stuff that’s un-necessary for a modern mobile web experience.

How did it go?

Remarkably well but there were some learnings along the way that are worth passing on.

At first, we were not as clear as we could have been about required functionality. It was all in Den’s head with very little on digital paper. That meant the design brief started out a tad fuzzy. As we got into the details, we found we needed to be specific on things that for us had become obvious. What we didn’t initially realize is that what’s obvious to us isn’t necessarily obvious to others. In short we made assumptions that were not well communicated at the start. Fortunately, Devrix were very patient and led us down the ‘right’ path to getting what we need.

One thing we did very well was dividing tasks and letting people get on with things. For example, Phil and Jon led the complex cornerstone part of the project and everyone else kept out of the way until they were ready for us to do our curation work. The team was also good at letting the devs get a good way down the track before creating a ‘fix’ list. This was important because we needed to ‘see’ the whole thing before making a start on change requests. That list became golden as we got close to the finishing line and were looking at many moving parts running in parallel.

As always, there were a few bumps along the way. Some things we would have liked to get done one way had to be iterated another way. That’s the world of constraints all technology lives in and with which you just have to compromise. Getting precious about things gets you nowhere.

We recommend fast decision taking. The devs came to us with ideas that on occasion challenged our established view of what diginomica should look like and what it should do. The temptation is to let things settle and mull over decisions. We preferred to get things done on the basis that if it made immediate sense then get it done and move on. We didn’t labor over any specific point for more than a day, although some issues needed iteratively hashing out.

Another thing we recommend – user feedback. Late in the day we were faced with an important design decision upon which we could not all agree. The answer came from polling 1,400 of our most loyal readers. We got a 56% response rate and went with the majority vote. That makes diginomica a reader’s site, which is how it should be.

Finally, we got this done on budget and with very little time delay given the project spanned big holiday seasons. That’s a huge bonus.

Where to from here?

We are far from finished. Diginomica is a dynamic site and we plan to keep moving the ball forward.

We’ll spend the next few days ironing out nits and then move on to the next stage which is to transfer the whole thing over to a managed AWS infrastructure. Assuming that goes well then we will have options to optimize performance in ways that are impossible with our existing infrastructure providers. That should mean improved performance anywhere in the world.

There’s a lot of things we can now do that were impossible with the old site. Most important is that we can personalize content for readers across 18 topic heads. That’s a huge benefit for anyone who likes what we do but only wants to consume sections of the whole content.

Now that we’re on the personalization path, we can augment content in many different ways. We haven’t figured out exactly what that means but we will, based on what we learn, our partner’s suggestions, how it works and what readers tell us.

Finally

Thanks to everyone who contributed to this project. It’s been an amazing journey and one which we can look back on with a sense of achievement that we could never have dreamed possible. Devrix did a fantastic job, working across nine timezones and, at the end, being available at ridiculous times of the night. Thanks especially to you, our readers. We would not be where we are today without your continued and enthusiastic support.

Disclosure - Oracle, SAP and Workday are premier partners at the time of writing

    Comments are closed.

    1. Redesign, looks good works well from mobile web thru desktop web .

      The font has higher contrast for my eyes (important in changing mobile light conditions).

      Cornerstone pages – choose your topic and community, subscribe and prosper. There’s a way to move cornerstone pages to the next level (after full migration to AWS). Use machine learning to build a personal recommender for each person revisiting Diginomica.

      Use case.
      Clive lands on Diginomica and visits Phil on Xero. The ML classifier on the AWS ranks all Diginomica articles in last 15 days into a custom sort and paints the page while clive is reading Phil’s and presents this personalized page for continued reading.

      Further columns can be instantiated if clive searches prior to recent 15 days (reducing cloud compute expense to already cached articles).

      To persuade clive he should sign up for personlized Digimonica pages a column can show a sneak preview of upcoming articles (which may change) by Den, Jon, Phil etc (could do so with articles over 15 days).

      A coda. FT.com includes a recommender list as does Digimnica for comments, make this personalisation more central to design. This can highly relevant personlized page clive could email to him self for later reading.

      See zulily android app (not main site) for an approximate working version of data science for media

      Btw, for frugal ML techniques to solve the cold start problem for new unregistered visitors see http://www.h2o.ai/docs/