SAP Analytics Cloud anno 2018: the good, the bad & the ugly

With all the updates SAP Analytics Cloud (SAC) got in 2018 (22 waves!), you’d almost forget that it is just 10 months ago that SAC was pushed forward as the future of BI for SAP. But the question is of course: can it already deliver on this promise? And is it really enterprise ready? So let’s have a look at the state of SAP Analytics Cloud anno 2018. I grouped my observations in a good, bad and ugly manner.

The Good

Let’s start with some important signs in addition to the famous strategy blog of February 2018. SAP Analytics Cloud will be turned into the single BI solution for all tooling within SAP. So not only as the front-end solution on top of a SAP BW or HANA system, but it will replace the built-in BI solutions that come with the several cloud solutions SAP has acquired/developed over the years. Think of SuccessFactors, Concur, Fieldglass, Cloud for Customer. This will bring in millions of potential new users. For project Blueberry, which is a future SAP BW4/HANA cloud based offering, SAC will be the BI front-end. For the on-premise tools the future looks pretty sad. They probably will be supported for another decade, but all new innovations will be done in SAC.

I also see that SAC is slowly transitioning from a BI/planning tool into a broad platform that goes further than what the SAP BusinessObjects BI Platform ever delivered. Initially it only hosted the models and stories for reporting and planning, then the boardrooms and predictive scenarios were added as separate objects, and next year we will see applications, Roambi templates and Analysis Office workbooks appearing.

The Analysis Office workbooks are an interesting topic in itself and I regard this as a must have for 2019 to completely get SAC to a full “enterprise ready platform” level. Currently, Analysis Office workbooks can only consume acquired data models in SAC. To make this really work we need support for live BW/HANA models, and the option to store these workbooks on the SAC platform from within the Analysis Office software. This is all on the roadmap for 2019.

In 2018 a lot of generic SAC features were released, for example story bookmarks, offline/live data blending and proper platform usage statistics. Also many more specific features were delivered, for example on BW live connectivity where SAC is now very close to supporting the full spectrum of BEx features. This year brought the universal display hierarchy, support for two BEx structures and support for BEx conditions.

To finish the Good section, SAP is doing a lot to support organizations to manage this new normal of cloud. It will offer so-called test tenants that run a preview of the upcoming wave. This allows customers to check out new features so they are prepared for the actual release.

The Bad

A lot of obvious stuff is still missing in SAP Analytics Cloud: we are for example still waiting for an undo/redo option. There is also still a huge feature gap between using acquired data versus live data, and even live BW versus live HANA. This makes it hard to learn the tool, as it is not always clear in which scenario you can expect a certain option to be available. The custom formula calculations are quite limited compared to what a Webi offers and the SAC platform lacks options for report bursting and alerting.

Also on the administrative side a lot needs to happen. User management is completely isolated from BW, BOBJ or HANA systems. You could use SAML attributes to automatically map accounts to certain roles, but the teams still need a manual assignment. Setting up and maintaining folder sharing rights is a hell of a job and very error prone. It is clear that the current concept was setup years ago for small use cases and not for an enterprise wide usage with hundreds of public folders and thousands of users. This really needs to improve as soon as possible; just take the BOBJ approach as an example I’d say. We can expect a first step in this in the first quarter of 2019 when the models will move to folders and start using team assignments for authorization, instead of the roles in which they currently need to be maintained. To end this part on admin stuff: the deployment workflow (lifecycle management) is still very chaotic and incomplete (not all objects can be transported).

Performance is still an issue in some cases, although really a lot has improved over the year. In a SAC story, each widget (chart, table, input control) is linked to a model which contains the data or brings in the data from a live data source. Unfortunately, even if you use the same model for multiple widgets, each widget will trigger its own query and impact the data source and SAC running in your browser. Furthermore there are no options available to influence the order of loading with a story page. An important recent improvement in this area is that input controls aren’t completely loaded upfront anymore, which improves the startup time significantly.

For the SAP Analytics Hub not much seemed to have happened this year. There were only a few waves and the only major change I can remember is the option to make assets visible to certain teams only. This is by the way completely in contradiction with the original philosophy of the tool, which was to present the users with allthe available (reporting) assets within your organization. Anyway, it basically is still the hub from 2017 when it was introduced: customization options are very limited, you can only define and use a single asset template, not all of the asset text is searchable, there is no real integration with BOBJ or even SAC (to select a story for example) and it still has a single tenant, production-only approach (no deployment). Finally, I’d like to see the Hub as the real start page of SAC, or at least give us the option to get rid of those awful default sample story tiles on the home screen.

Moving to mobile. It is almost 2019 and there are no signs yet of an SAC app for Android devices. Even as a lifelong Apple user this worries me. Another point of concern is SAP’s strategy on app integration, or better, the lack of it: Roambi templates are about to be integrated in SAC, but the apps remain separate! So that is two different apps to connect to the same tool. Even worse, neither of these apps support the SAP Analytics Hub. So you’d need your mobile Safari or Chrome browser to go to the Hub first and jump to a report in one of the apps from there. The Hub site is nicely responsive so this actually works very well on an iPhone or iPad, but you really don’t want to go back and forth between multiple different apps all the time to run your reports.

Finally, we saw the dark side of SAAS a few times, when there were some major outages in the German data center that runs SAP Analytics Cloud.

The Ugly

I’ve never been a big fan of the model/story concept. For acquisition scenarios and planning I see the added value, but for live data sources I’d rather just select a BEx Query of HANA view directly from the story. Unfortunately the concept seems to remain for all the SAC objects, although there are some ideas on offering a “generic analysis template”-like solution within SAC, but it is completely unclear what this would look like and when it will be available.

Another concept that I don’t see is the story/application split. Early next year Application Design will be added to SAP Analytics Cloud. You could see this as a Lumira Designer in the cloud solution. Instead of completely integrating the application design options in a story, the decision was made to split the two and create a separate “application” object in SAC. The problem that I have with this is that you have to decide upfront if you should go for a story or an application. You cannot extent a story with Application Design capabilities (e.g. scripting) or vice versa. Good thing here is that SAP is thinking about a one-time migration option to go from a story to an app.

Another ugly one is the following: To make the feature gap between the different (live) data sources even more complex, a lot of interesting new features are only supported in HANA 2.0 and won’t be ported to HANA 1.x. Think of blending with acquired models, calculated dimensions and measure based filters.

Finally, SAP moved from a bi-weekly release schedule to a quarterly release schedule. New customers will be automatically on the quarterly schedule. Ugh.

Conclusion

2018 was a great year for SAP Analytics Cloud and we’ve seen some huge steps in the right direction. The reporting and planning features of the tool are solid and have been proven in serious environments. The unique combination of reporting, planning and predictive capabilities, with the option to create advanced custom applications in a cloud based platform that integrates very well within SAP and non-SAP data sources is of course gold. It has all the potential to be something different, and better, than other tools in the BI spectrum. As you’ve seen, my wish list for SAC is still large, but nothing that can’t be fixed: the roadmap and future direction look very very good.