The Seven Samurai of Domo.
We have far too much data, and it’s too important, to keep it on premises.
- So said a customer, untangling herself from a ski fitting, at one of the Domopalooza 2018 evening receptions. Cloud computing has come a long way in recent years, in capacity, affordability and trust.
- In this second of three reports on Domo’s annual customer gathering, I’ll focus on technology news from the event. As in my first blog about the customers I met, rather than provide a full of overview of Domo, we’ll talk about what’s new and what I specifically heard at the conference.
The Domo dojo
- The teams at Domo clearly like to do things differently, with an aptitude for branding and style, and a reluctance to lead with technical messages. In short, there was no marketecture: none of those high-level schematics of boxes and arrows, which vendors use to oversimplify their system components, layering a favorable gloss on what may be a messy reality under the hood.
- Domo, in their distinctive way, avoid that traditional route altogether. They don’t even try to map out an architecture, but prefer, with their somewhat mannered orientalism, to characterize their underlying capabilities as Seven Samurai. Each member of the dojo is itself a construct of often complex components and features, many of which may overlap. It’s probably easier to show you what I mean. Here are the seven:
I’ll not go into all of these in detail: a couple of examples of Domo’s characterization will suffice for now.
- Mr. Roboto: an automated intelligence platform, introduced last year, of artificial intelligence engines, including machine learning, that constantly traverse incoming data and behaviors to detect trends, anomalies and correlations, providing alerts and facilitating business processes. With Mr. Roboto, customers are guided to areas of the business that need their attention and are freed from spending time on areas that are running according to plan.
- Adrenaline: an exo-scale, massively parallel, and sub-second deep learning Data Warehouse engine. Adrenaline also includes Vault, a next-generation data lake that serves as the full, comprehensive data store of record for many Domo customers. With Adrenaline, customers have an intelligent data warehouse that can be leveraged to actively drive more insights and value from their data.
- You get the idea. There’s no significant technical detail here, but some big claims. It would be tempting to try to match up Domo’s Seven Samurai to the original movie: Adrenaline is perhaps Yoshio Inaba’s brilliant archer and strategist. Mr Roboto could be Mifune’s committed but volatile fighter who has the brains and skills, but not the noble background of the others.
- Let’s not push the analogy too far … and let’s be thankful that Domo’s marketing team have not based their model on the 47 Ronin. Seven is enough.
- Domo’s habit of going with quirky names strikes me as sometimes more confusing than helpful. When they talk about Beast Mode, they mean little more than calculated fields with a SQL-like syntax. On the other hand, this approach does build a community of practice with an almost tribal allegiance to their own arcana. And it leaves the architectural dialog of other vendors sounding sterile.
- Speaking with the tribe – the Domo users at the event – they mostly like this approach. It doesn’t affect their belief in the technology. One system architect who delivers BI infrastructure to a few hundred business users, told me, with a certain wonder, that they had quickly adopted the dramatized terminology of Connect, Fusion and Adrenaline making for more natural business conversations than users having to adopt the quasi-technical language of ETL or data warehouses. What I heard at the event confirmed this to some extent: most users spoke about Domo’s architecture in terms of the Seven Samurai, fluently and comfortably.
What did Domo announce?
- It would be a mistake – and easily made – to assume that this quirkiness masks a shallow attitude. However, several of the product announcements at Domopalooza showed product teams tackling data governance and application management issues with a serious design intent.
- For example, administrators can certify data sets, calculations and visualizations as trusted sources. Users can add business metadata to data sources to describe the business context in which the data should be used, and the automated analysis features of that most engaging Samurai, Mr Roboto, provide data profiles for assessing quality and completeness. Of course, similar features are found in numerous other data platform tools, but that’s rather the point. The prominence given to connectivity, data management features and built-in dashboards for system statistics in Domo’s announcements shows clearly enough that customers increasingly see Domo as a data platform, not just a front-end analytic tool. The new marketing message rolled out at the event shores up this distinction by calling the platform the operating system for your business, with the implication that there’s nothing to worry about but the business.
- Reviewing Domo’s visual design features, I was rather taken with Pixie – another feature with some serious technology hiding behind a cute name. Pixie is an automated dashboard layout tool that uses prebuilt design templates and some smart understanding of the underlying data to map out visualization components in ways which are optimized for mobile, web, printing and so on. I have seen – and prototyped! – this kind of feature many times, and it can prove to be a complex business. The demo looked excellent, but the devil will be in the detail. If Pixie can generate useful layouts first time in the majority of cases, users will be released from some drudgery: if the results mostly need tweaking and customizing, the novelty will quickly pall.
- Naturally, Domo have built-in some ability to customize Pixie’s layout options. Similarly, there were several other announcements: customized inclusions in the Domo workflow and analytic capabilities: custom charts enable special visualizations to be used alongside Domo’s standard cards, alerts can be personalized with dynamically generated messages, R and Python modules can be integrated into Mr Roboto’s workflow. All good signs of both a flexible underlying architecture and, what I suspect is an increasingly open attitude at the front end.
The feature I wish I had built
- Speaking of Mr Roboto, he not only has the cool name, he gets the cool feature announcements too. The most significant for business users – and I expect a powerful differentiator if Domo can keep it simple and message it well – is the ability to add predictive capabilities to almost any card or data set.
- I have wanted to build this functionality into my own products for years. However, it has not been until recently that two important elements could come together to make this happen. Firstly, developers need techniques to quickly build and train an appropriate model over the data without specialist intervention from the user. Secondly, business users need to be data literate enough to make use of the feature without training and without having to work too hard at interpretation.
- Imagine a supermarket manager, reviewing weekly store sales and noticing some items of concern, especially the poor performance of special offers in some departments. They want to track new offers in real time and know if they need to intervene with more promotion, or repositioning the sale items, or some other action. With Domo, they can easily create a predictive model over the real-time data being collected in the store and set alerts on the outcome of that model. The managers understanding is sharpened when the model predicts that the offer will fall short of target, rather than waiting to review the offer at the end of the week. There are endless variations on this scenario which will be powerful and valuable.
- Chris Willis, Domo’s thoughtful and erudite VP of Design demonstrated this feature very effectively. Success on the road ahead will lie in how easily users can configure this feature and how reasonable the results will be to a non-specialist business user. Chris made it look easy, of course. Users I spoke with at the event were excited by the concept, and had some confidence that Domo could make it work for them.
The bittersweet ending
- Domo’s discourse – the conversations they lead with customers and analysts, their form of messaging – not only leads with summary discussions about capabilities and business use cases, it tends to stay there. At other vendor events, engines and stores like Adrenaline or Vault would have been treated to a dress parade of technical detail.
- If that’s what you look for from a vendor, then Domo’s withdrawal from the common language of the BI market might feel perversely shallow or evasive and at worst literally mendacious.
- That temperament has changed, but there’s still an undercurrent in Domo’s conversation that suggests we should just trust them on the technologies and be awed by the customer stories. Even during the analyst break-outs at Domopalooza, we had a discussion under non-disclosure that was very light on detail. This is so different from the approach of other vendors that it can be quite disconcerting, especially if, as an analyst, you want to fit Domo into a predefined model of what business intelligence should be, what the challenges are, and how the software market operates in that space. Domo consciously and resolutely do not play the game by those rules.
- Yet for now, and rather to my own surprise, I’ll give Domo the benefit of the doubt. For two reasons.
- Firstly, I have had no problem getting technical details from Domo when I ask. For every question I have posed, I have been connected to engineering team members for unfailingly well-informed and open discussions. I have enjoyed long conversations about technologies underpinning the platform and design principles that have been as informative and insightful as I could want.
- Secondly, Domo’s unworried customers clearly endorse this approach. They see a potency in the Domo story and like the way it is told. Ultimately, that’s the endorsement Domo wants. At the end of Kurosawa’s movie, when the farmers have been saved from the bandits, the leader of the surviving samurai reflects that The victory belongs to the villagers, not to us.