SaaS Applications, Integration and Business Process

Recently VentureBeat wrote a post that discussed fragmentation between SaaS solutions and the impact that has on organizations using different services. It is a topic that I’m interested in, not least of all because one of the businesses I’m involved with, a manufacturer and distributor of backpacks and apparel, has long had to wrangle disparate software in its operations. In the post the story was told of two businesses – a spa installation business and a business travel booking service – two very different organizations but both of whom feel the pain of getting discrete applications to work together.

The post pointed to a Forrester survey taken this year which found that 64 percent of midsize companies are planning to focus on improving their workflows this year. The key point here, and one which I was pleased to see, was that the post articulated the problem in terms of workflow, rather than in terms of integration. As I’ve pointed out before, businesses are looking to the cloud for their solutions but, while individual solutions might fulfill their needs for a particular problem, there is little focus on the wider business process itself – integrations take a very micro view of taking data from point A to point B rather than the more macro view of fulfilling the needs of the wider process.

I was pleased to read the comments of Jason Lemkin, founder of EchoSign and one of the sharpest guys in the SaaS space who reiterated a theme I’ve grown fond of articulating:

We need SaaS startups that automate much of what is still manual data-entry in current generation SaaS products. Software is great. Web software is better. Data entry is horrible. Data entry by high-priced employees who aren’t paid to do data entry? Unconscionable

We’re seeing a generation of applications commoditize and greatly increase the availability of simple data integration – IFTTT and Zapier are doing a great job of this “plug and play” integration platform opportunity. But it’s wrong to mistake data integration for true workflow enablement. As I said in a comment to the original post:

Integration is step one in the process – but only a very small step. Data integration without really looking at the real problems for small and mid sized business is counter-productive. Integration companies need to look at the very real issues around workflow differences and user experience – this is the critical factor and the reason that SaaS integration in and of itself isn’t enough. Focus on UX, not data…

That’s a lofty goal but the difficulty is that pure data integration is a fairly simple concept – “I want data A generated in application 1 to correspond to data type B in applications 2 and 3″. True there is some plumbing needed to achieve that but essentially it’s about creating some pipe work and enabling data to move between applications. Workflow and UI integrations are very different – partially because they relate to not only the data tier of the application, but also the presentation tier. Another complicating factor is that UI is so specific to an individual situation. An accounts clerk might have an ideal offering as one which displays CRM data in their ERP application for example, whereas a CEO might want a generic dashboard applicable to all the different application types in use in her organization.

Possibly the best solution we can hope for is some sort of neutral abstraction that takes two different data types and presents them in an entirely different way – true this create a bit of a learning curve that users need to get over, but it has the benefit of being able to deliver a UI specifically tailored to deliver the best experience for the combined data. This is the approach that WorkinBox is taking (disclosure – I’m an adviser and investor) – taking email and CRM data and presenting it in a new way – tailored to the combination. Other approaches are like those taken by Zapier which is a more menu based approach – it’s still primarily about data integration but as workflow processes start to be offered on the platform that will change I believe.

Possibly the closest metaphor for the ideal integration, workflow and UI offering that I’ve seen came from Yahoo! a few gears with their Yahoo! Pipes product. Pipes created a visual interface where users could combine different data types and build simple (and sometimes not so simple) workflows with those data. The outputs could then be introduced into different applications or presentation tiers.

This is very much a nascent space and one which I suggest we’ll see much development in over the coming years – but it would seem to me that any SaaS application that has an inflexible and one dimensional UI and workflow process is ultimately doomed to be disrupted by more open, flexible and user defined approaches.

1 Comment
  • Dear Ben, nice article but as an integration platform for data integration between SaaS (elastic.io) I would like to add some other elements to the discussion:
    a) UX integrations alla Eclipse (in the on-premise world) are difficult to realize as most SaaS do not allow access via API only but require the user to register on the website
    b) UX integrations cost a lot of money; most SaaS users are SMEs which do not budget for customizing which goes to expensive consulting resources

    So, data integrations are a good start until there are more backend-as-a-Service available.

    We believe the integration should start where the problem occurs which is at the SaaS website. Thus, we earn money by providing SaaS integrations as a white-label to SaaS companies. The underlying idea: SaaS need to offer their end customers data integration left and right of their particular vertical as their end customers think in horizontal business processes. This way, the SaaS maintains and improves its end customer relationship and gains new ammunition to attracting more users. On our platform the SaaS constructs its use cases and embeds them in its website – a plug&play integration App Store filled in minutes instead of days whilst elastic.io stays in the background. The end user activates a use with just one-click – no more fine-tuning required. The best thing: It is all priced as-a-Service (1 € / end user / month), so no risk & scaling!

Leave a Reply