Top-of-the-market closed software vendors may become seduced by their own power and turn complacent.” – Sergei @ Daxtra
I can’t believe I’m saying this but…
It just stuck in the throat, but after 10 years of developing a business whose fundamental job is integrating systems (we post advertising data into 1200+ job boards globally) – I find the technically filthy iFrames address certain issues in the world of integrating systems that are good.
Let’s come back to that…
The difference between open and closed systems is explained well by Sergei Makhmodov in this article – and while closed systems, controlled by the vendors served them a purpose historically, I’m happy to see that it would be the undoing of any Recruitment System moving forward in our SaaSsy world. Good.
Legacy big-boy market-leaders have done everything possible to not open up – or to do so on the basis of +$1600 day rates with a 3 month wait time. Removing their control not only allows their client to become more effective as a business, its good for the market/innovation/competition in general.
In the new obsession to integrate systems (think Zapier, IFTTT, Internet of Things etc) it amuses me to see how what we saw as historical technical inferiority becomes flavour of the month.
Webhooks are now the darling of the SaaS world – they allow any system to push data to any system that can accept it (note, only dinosaurs do not – we should be releasing our idibu web hook system in the next few weeks).
When we first started idibu we learned about all the different ways to integrate and send data into job boards – the worst approach was plain old email and FTP (you can hardly call this an integration frankly) – and the utopian vision was ‘the official API’.
Well, just above FTP was what idibu called “HTTP Push” – it’s where you post all the job details in the URL – to inject the job advert. It was deemed a slightly dirty solution, it meant the job board didn’t need to develop a technically robust API… and most of the time it worked… most of the time.
Fast forward 10 years and the world is in love with Webhooks – which are simply “HTTP Pushes” with a new name – no longer dirty hacks – they lower the barrier to entry, and enable semi-technical people to quickly hook together disparate systems. They are great. And they haven’t just achieved main stream acceptance through marketing fluff – the network infrastructure connecting all our system has of course improved too in the last decade.
Of the difficult lessons learnt from integrating idibu into CRM/ATS’s over the past 10 years, you would think the purist API integration route would be the way forward – I know I did.
The problem is that you end up with a huge number of different systems and user interfaces – sometimes created by a partner who ignores your advice on best practise – which means your support team have a scalability issue as you try to deal with all the nuances of these different models. Yes, a full API integration is theoretically the best. In practise the freedom to ‘build it as you want’ can come back to haunt you.
Yes, a full API integration is theoretically the best. In practise the freedom to ‘build it as you want’ can come back to haunt you.
Limitation can be a liberation – and in planning how we would integrate the new idibu platform into Bullhorn we had to grapple with using a teeny-weeny iFrame window in their system.
This turned into a blessing – we re-engineered our navigation completely, upgraded our core system to work effectively in this small space – in short, by accepting the gross limitation of an iFrame integration it forced us to push what we could do.
It’s made us really work on the user experience and to focus acutely on helping a consultant use our product from inside their CRM.
And the best bit?
I love the simplicity of this new low integration barrier to entry where our service can be fully embedded inside any other system in hours. I love that Webhooks let us send key events to any system anywhere instantly.
Sometimes the best solution to a problem already exists – and producing new technology is not the answer. Improving what already exists can often be the non-revolutionary but excellent way forward.
No Comments Yet
Let us know what you think