Key considerations when shopping for a hybrid integration platform
When it comes to Hybrid Integration Platforms (HIP), I like what Massimo Pezzini and Elizabeth Golluscio have done with defining a 4D view of the greater Hybrid Integration context. Massimo is correct in stating that too often the focus is on the deployment model, and not on the users, the integration patterns, and endpoints you need to interact with.
But I think the recommendations fall short in providing real guidance beyond building this mental model of the problem space. My goal here is to try to continue the discussion Massimo has created with some tangible next steps to consider, and more importantly, some imperatives you should establish along the way:
Imperative #1: Are we for, or against, Shadow IT?
This is an important question to ask yourself because it has huge downstream ramifications. If you’re for shadow IT, you likely see the need to go fast, and you like to let lines of business try out best-of-breed solutions. If you’re against shadow IT, you want to ensure integrity with your data and security practices, and a general priority on governance.
These are important decisions because it may mean the difference of giving marketing $500k to spend on the latest and greatest customer segmentation analytics tools with built-in integration, or IT spending $500k on CASBs to provide the governance mandated by the business. Sit down, have the debate, decide as a business. Your Hybrid Integration Platform decisions need to take this into account.
Imperative #2: Map out your 10-year integration needs
Yes, that sounds absolutely like an insane exercise. Sigh, I know – the landscape will change, there will be disruption, none of this is real, … It’s a planning exercise, so don’t overthink it.
Don’t try to predict disruptive technology here or anything. Think of it in your terms, based on how fast your company can (or should) move. Do you need a marketing data lake to get deeper customer insight? Am I building a new order entry system to support online b2b efforts? Am I doing location-based marketing campaigns that need to act quickly?
Think about what you’d like to implement in the next 10 years if things went smoothly. Now, rate your integration vendor against those needs. Chance are you’ll find you need a platform that does some Big Data things, some real-time things, some process things, some analytics things, … Use-cases are converging, and so integration is converging. Are you sure you don’t need to orchestrate a process of machine learned actors subsisting on a diet of big data and real-time streams in order to automate your suppliers?
The key here is, don’t rate a platform decision on the urgent issues of today. Platforms take time to get right and they carry technical debts forward a long time. If they fail to expand beyond initial use-case after 2-3 years, their risk of death is high. Also, see Imperative #3.
That’s what marriage does to a man’s analogies I suppose, but the idea is, is this a throw-away integration or a long-term investment? Fast fashion, or, less but better? There are a set of simple integration scenarios that can be accomplished very quickly with simple SaaS services for a low monthly fee. Or basic data integration needs that are better done by a Data Prep tool.
These same tasks might otherwise take you 6 months with an ETL tool or ESB. The whole project might live for 18 months. (It creates a concept I find very funny: In technology, litter is useless code that is still being serviced, but throwing stuff away is good because it eliminates litter.)
On the other hand, a new mobile order entry system with image recognition and fingerprint verification will probably be a long term investment, and one that you absolutely need to get right. Doing it slower but better is a worthy tradeoff you want to make, and over the course of the integration it will more than pay for itself.
So when you consider your 10-year needs, think about dividing them not only on short-term/long-term, but on the tactical vs. strategic. Forever 21 or Cuyana.