This week was tiring in the extreme – but in a good way. And tomorrow morning, we are leaving on vacation to Hawaii. The plan is to switch off work email and phone and just chill with my family.
Last year – we went to another island in Hawaii around the same time. The week before that vacation was also very stressful – but in a bad way. Something that I was working on with all my energy did not come through – and I decided to take a vacation to get over it. We decided on Friday, and we left for Hawaii the next day. It was awesome – and I came back fully refreshed, and the thing that did not materialize before I left, that happened in few weeks.
This time is a little different – we hit the results we wanted with the launch (http://www.saphana.com/community/blogs/blog/2013/09/24/announcing-sap-bw-on-hana-trial-offering-on-sap-hana-marketplace#comment-4174) , with room to spare. And I am leaving it in the capable hands of my main man Ingo Hilgefort while I am away.
As the last meeting of the week got over, I have been thinking through our little side project. What immedeately striked me was that we never bothered to give it a code name. And it never mattered once – we all worked as hard as we could without slogans and call names. I now need to think if this should be the norm going forward for my gang.
We did this with a core team of three – Ingo, Rohit Kamath and me. Ingo acted as the project manager. We had a lot of other things to do in our day jobs – so this was a side project with no formal budget of headcount. We went right ahead – and pulled in colleagues as needed from every corner of the SAP organization globally. And no one hesitated to help – every colleague we asked help, turned around and helped us. Of course some more so than others. But no one turned us down. As it goes – ask and you will be given .
Why did we do this? Because we were in front of a lot of customers and partners across the world in the first six months of the year. We knew first hand that it is painful for ecosystem to do POCs – and people told us versions of the same story in every meeting and event we had. So it was a no brainer for us that this needed to be solved.
It also helped that Vishal challenged us to think of a way to get a POC system up and running in 3 days, instead of the weeks and months it typically takes in many projects.
It was not difficult to arrive at a conclusion that cloud deployment is the best way out. The big question then became – will BW shops use cloud at all? We all had different opinions. So I wrote a blog about it and put a poll inside it to see how the ecosystem percieves BW on Hana on cloud. The data – although a small sample – clearly showed it was a welcome idea. So we went ahead with the plan.
Next up was the question of what is needed for the POC system – and after some back and forth with ecosystem – it was agreed that BW alone didnt make sense, and that we needed BW on Hana with a BI server so that customers can better visualize the awesome capabilities of BW on Hana.
This solved half the problem – the other half was more difficult. How do we ensure that a customer will face minimum fuss in doing a POC? We analyzed further and figured out that it is not enough to provide raw infrastructure – we should also provide connectivity between BW and BI, as well as sample content and data. And for those who need more training – we also decided to provide tutorials. This is the same material we had tested with probably a thousand people before in our enablement sessions in the first half of the year. So we were sure that this was the best way to get someone up to speed.
Then came the actual deployment. We had many options – many different clouds, many different stores, many different integration technologies etc. We decided to make SAP HANA Marketplace as the entry point, make SAP store the transaction engine (record keeping) , SAP CAL (cloud appliance library that does the magic behind the provisioning) as the deployment engine , and AWS as the deployment infrastructure. Together – they offered the best trade offs.
At that point – we needed the best critics we could get to see if our plans hold water. That is where the SAP Mentors threw us some help. They got into a call with us and gave us excellent feedback. And to fill the remaining gaps – I also ran it with Dennis Howlett who also was helpful in his feedback. We listened and we acted – and that is what eventually went into production on September 25th morning.
We only did a soft launch – Ingo and I wrote a blog each, and we tweeted out from our personal handles. It was not exactly a social expiriment – it was a well thought through plan with colleagues in marketing. The results surprised me to no end. Thousands of hits on the blogs, plenty of twitter amplification, and significant traffic to marketplace. Best of all, way more people who reached the market place finished the transaction and got trial systems than we expected. I was skeptical that the numbers were any good – but after seeing comparables from people who do this thing for a living, I think we did pretty well.
You will see a lot more activity from next week to get the word out across all channels. I cannot thank enough everyone who is working hard now to get this in front of as many customers and partners as possible. It is also amazing how many colleagues pinged us directly to ask how they can help. That is what makes it special – where people volunteer to help when it is not their day job.
From the time we launched – we have been staying in close touch on all channels to provide assistance to people using it. Where we cannot help – like for example, there are only a subset of countries where we can legally offer this today – we are updating the marketplace with information about what is available and what is not. I am especially thankful to everyone who tried it out – and then took the trouble to communicate what they were happy with and where we need to do better.
There are some 18000 BW installations today – and BW on Hana is something all of them should try out. This is our first step to make it very easy to do so. Even more important, the feedback we get from this trial is what will drive our plans for what else we can do. Our hope is that data will tell us what they next step is, instead of us guessing what customers need.
It was a learning experience for Rohit, Ingo and I – and probably the 40 or so other colleagues who helped in the extended team. It gave us all a better idea of how many things are interconnected in a big company, and how we can all work together in informal settings to make good things happen. Sure there were many frustrations and crises along the way – but we prevailed as a team and got through it all. I am incredibly proud to be a part of this team and will gladly do it again. We also learned that we have to take turns in calming each other down
What would I change in how we did this?
1. Now that I know how many people have the kindness of heart at SAP to help – I will crowdsource more, and start involving more people early on. If that needs these things to have code names – so be it :). I should have know this after the success I had with it creating the IBM innovation lab for SAP in a crowdsourced model with Gagan. Good to know this works in all companies.
2. A small core team is the reason why we could move in an agile fashion. I will keep that philosophy. Rigid structures kill initiative in smart people.
3. Social is definitely the way to go to push the message and get feedback. I think we will do way more social interactions in next project. Not just more social – we will start doing this sooner too.
4. Start setting up measurement systems sooner in the project – even if it is informal projects. Sooner the better.
That is pretty much it . Couple more hours to finish the day off – and then get busy packing. I have promised my wife and daughter that I will not think of work when we are on vacation. Let me see how much I can keep that promise.
Ingo and Rohit – you guys rock ! Cant say enough thanks to you – and all the other colleagues who helped and continue to help. Also a very special thanks to Amit Sinha , Jonathan Becher, Steve Lucas, Prakash Darji and Vishal Sikka (and their teams of course). We could not have done it without you.
Hope that by the time I return to work – I will see a lot more people using the trial. If not – we will go back to more analysis and fix it. Long complex POCs suffering from complexity of IT systems is a problem the world doesnt need – we will fix it some how, no matter what.