Every year, Mary Meeker – partner at Kleiner Perkins Caufield & Byers – releases her ‘Internet Trends Report’. It is probably the most eagerly awaited slide deck of the year (alongside Stacey Harris’ HR Systems Survey and Research report – which if you haven’t yet filled it in, you can do so here). Mary’s Internet Trends report gives a facts and figures based summary of the current state of the marketplace, and predicts how that’ll change over the coming years. It’s lengthy, but always a treasure trove of information.
The slide deck can be found here, however I’m going to highlight a few items of interest and point out how they’re relevant in the world of Oracle.
Mobile vs Internet use
Both mobile and internet use are still on the rise, however the increase in mobile handset shipments is decelerating. Are we approaching saturation point with mobiles? There are 3.4Bn internet users – approx half of the World’s population – and this is increasing by 10% each year, with the number of mobile handsets shipped at just below 1.5Bn – with more than 80% of those Android.
What’s also intriguing is the number of hours spent each day using your mobile is now averaging 3.1 hours (up from 0.3 hours 8 years ago), however this isn’t coming at the expense of Desktops/Laptops (which averaged 2.2 hours last year and 8 years ago). So the amount of time we spend interacting with digital media has increased hugely in that time.
Oracle has improved the responsiveness of the HCM Cloud in recent versions, and this is – rightly – going to continue in future releases. It’s important that the desktop interface is still improved at the same rate, as the figures show that isn’t declining.
The report states that 20% of searches are now performed by voice command, rather than text input. In the age of Siri/Google Assistant/Alexa this should be no surprise, but it does illustrate that we’re on a path towards voice interfaces being more commonplace. The accuracy of systems understanding voice commands is also on the increase – it’s now around 95% (which is about the same as humans).
Clearly chatbots (whether interacted with via voice or by typing text) are the future for much of HR Self Service interactions.
Doormen = Foremen
We were tickled by the pictures of doormen in apartment blocks becoming more like foremen and needing a room to use as a warehouse for holding all the parcel deliveries that come in via the post for residents each day.
Lowe’s Augmented Reality app to provide in-store directions to goods on their shelves is very cool.
Augmented/Virtual Reality is going to be important in the field of learning. Most people learn better by ‘doing’, so a situation where trainees can perform a task in a virtual environment – free of consequences if they make mistakes while learning – is a useful tool. I’m not sure we’ll see AR/VR in HCM Cloud any time soon though.
Instant Gratification and Gaming
Millennials are growing up surrounded by gamification, and apps and games that give instant feedback by ‘levelling up’ quickly. Different games are preferred by the generations, but many of the mechanics that drive players to return are consistent across the games. We’ve all seen sites that rank your profile as XX% complete, but what is newer is the rise of services such as Twitch – where live games are streamed to interested viewers – which is considerably more popular amongst younger generations than older people. An example of this it the recent League of Legends World Championship received 43 million unique viewers.
Oracle Cloud Connect now highlights its top 10 users of the month, rewarding those who’ve put effort in to the platform. Instant feedback is also coming to the HCM Cloud, in the form of kudos now becoming real-time performance feedback, rather than the annual routine that we used to follow.
Developer / Designer Ratios
A short topic that caught my eye was the change in Designer vs Developer ratios over the last 6 years. Back then, companies such as Atlassian, IBM and LinkedIn reported ratios of 1:25, 1:72 and 1:11 respectively, with Developers far outnumbering Designers. The ratio has moved in significantly however, with companies now reporting a ratio of between 1:5 and 1:9.
Oracle’s Apps UX team have been at the forefront of designing the HCM Cloud user experience, and can be seen at most Oracle events evangelising the design paradigms that they’ve used in this process.
The Rise of Digital Healthcare
The number of companies that focus on personal health has grown very rapidly, but this growth isn’t just in the field of wearables, as one might expect. Companies offering health-related datasets to dramatically reduce the time needed for clinical trials (and hence speeding time-to-market), patient Electronic Health Records and tests for genetic disorders that you can take via the mail (e.g. 23andMe) are all areas of growth.
Oracle HCM Cloud has a whole module for this, My Wellness.
It’s a common practice to change the background colour of your environments to give a visual clue which is which. This helps reduce confusion and lessens the risk of entering test data into Production.
Whilst the core project team might easily remember which colour corresponds to which environment (e.g. blue=Prod, red=Dev, green=Test) it’s not so easy for more occasional users. There is a solution for this.
We’ve created some semi-transparent backgrounds that prominently feature the name of the environment. They’re semi-transparent so that your background colour of choice can still show through. Here’s what they look like:
You should be able to see the words Test and Testing on top of the green background. This gives less frequent users more certainty they’re logged on to the correct environment. I’ve chosen a green background colour here, but any (reasonably dark) colour can be chosen and the lettering will still show.
Here is the Development background image on top of a red background colour.
The images work with both Release 11 and 12, and can be downloaded here:
Remember, the lettering is semi-transparent so they might look like a blank image until you select a background colour behind them:
I’d be happy to create additional environment backgrounds if enough people show an interest.
Earlier this week Cedar held its Oracle Talent Acquisition Cloud event in Oracle’s London City office. We have a really strong team of UK-based Talent Acquisition specialists within our ranks:
Jo, Prachi, Marc and Mel are all experts in the various aspects of Oracle Talent Aquisition Cloud – Taleo Recruit, Sourcing and Onboarding – and Cedar has multiple implementations under our belt covering all of the above products, so we were able to put on a seminar with strong content.
I’ve put a post on the Cedar blog with lots of photos from the day here.
Both Oracle and Workday have recently announced figures for their Cloud application suites and it’s interesting to see how they both compare.
In its Q4 results Workday stated that they have 1,528 HCM customers with 136 of the Fortune 500, plus 320 financials customers. 70% of their HCM customers are live.
In its recent Analyst Briefing Oracle shared that they have 1,600+ HCM customers with 275+ of the Fortune 500, and 3000+ financials customers. 1,000 of their customers are live (giving a figure of 62.5%).
Here’s what that looks like in pictures:
(Taking Oracle’s figures of 1,600+ and 1,000+ as 1,600 and 1,000)
(The Fortune 500 is an annual list compiled and published by Fortune magazine that ranks 500 of the largest U.S. corporations by total revenue for their respective fiscal years. This chart assumes that no corporation has both HCM Cloud and Workday. It also takes Oracle’s 275+ figure as 275.)
(Taking Oracle’s figures of 3,000+ as 3,000)
Those of you who know me will know that I’m working at Cedar and we’re moving from PeopleSoft to the Oracle Cloud. I won’t talk about the company’s journey, but more how I’ve found the transition personally.
In brief, there’s a lot to learn.
Data Model knowledge
In the PeopleSoft world, once you know the table structures you could work significantly faster. Whether you were writing an SQR, an App Engine, an XMLP report, a migration script, an ad-hoc Query or a piece of page code the source was always the same – the PS tables.
In the Oracle Cloud world the tables are still there, although they’re closer aligned to the eBusiness Suite data model so they’re not immediately familiar to those from the PeopleSoft background. This isn’t the end of the story however, as the raw tables are only accessed some of the time (for instance, during a BIP report). If you’re doing an ad-hoc Query in OTBI then you’ll need to know the Subject Area structure and this is very different from the underlying tables. If you’re writing an HCM Extract then you’ll need to know the UE structures, and they’re different again.
Is this better than what we had in PeopleSoft? These abstractions are good in some respects, for instance OTBI Subject Areas are easier for end-users to pick up than the database tables for ad-hoc querying. But it does mean that for techs there are multiple different data models that you’ll need to learn.
I was never involved in a File-Based Loader migration – luckily it seems – but we’ve used the new HCM Data Loader tool a couple of times now, once for a migration and once for a mass update of existing data (annual salary uplifts). It’s actually quite nice, once you get the spreadsheets setup – and I had a couple of colleagues helping with that – and the data goes in pretty well.
Is this better than what we had in PeopleSoft? Yes, I think it probably is. Many times the code for PeopleSoft migrations was handwritten from scratch and would be thrown away after it was completed. With HDL there’s a lot of re-use and it’s probably quicker to put together too.
In the PeopleSoft world we have three levels of role-based security, the user profile, roles and permission lists, all managed from a handful of pages and covering both page access and data security.
In the Oracle Cloud world it’s a much larger model, with abstract roles, data roles, job roles duty roles, security profiles. Each is managed from a different set of pages in completely different areas of the system. The system is moving in the right direction, as the Security Console does bring some of this into a single place.
Is this better than what we had in PeopleSoft? I’m not sure. I don’t consider myself a security expert in the Cloud yet, and although I can see it’s a larger framework with more moving parts I’m not yet sure whether this is a better thing.
In the PeopleSoft world we had a very capable toolset and the access to change whatever we wanted. This could be dangerous as although something ‘could’ be changed, it wasn’t always the case that it ‘should’ be changed. With a little experience it was generally pretty easy to see between these cases however. This generally meant that whenever there was an issue it could be fixed by getting into the code and making some changes.
In the Oracle Cloud world it’s a similar story when it comes to typical tech tasks – migration, interfacing, etc, but it’s a different.answer when it comes to issues with the delivered system. If there is a problem then it’s often the case that you’re helpless to do anything other than raise an SR.
Is this better than what we had in PeopleSoft? I think time will tell here. Everyone is still learning to work together in this area. Our SRs could probably be more detailed in order to prevent a lot of to-and-fro between Oracle and implementation team, and the Oracle support teams are still learning as well, especially as the system is moving target with new releases every 6 months. In general Oracle Cloud fixes get delivered about the same speed as PeopleSoft ones (assuming the client is on 9.2 with Selective Adoption).
Is my life any easier in the Oracle Cloud world? Not yet, but that’s largely down to the amount of time I’ve spent working on the Cloud vs my PeopleSoft experience. I’m confident that I’ll improve – and the product keeps improving at breakneck pace – and I’ll be able to help many more customers make the move.
Cedar held its annual Oracle Day in Oracle’s City office last Thursday – it was actually our 6th year of running the event. The attendance was great – just short of 100 – and with 3 streams of content there was plenty to choose from. If you weren’t there, here’s the highlights of the sessions that I attended:
After the opening keynote from Dan Woolstone, Oracle’s James King and myself spoke to a full room on Top 10 Tips for a Successful HCM Cloud Implementation. There was a good mixture in the audience of clients already running in the Cloud and those at some stage of their journey, and we received some great questions.
Swiftly following on was Mike Everitt peeling back the curtains on HCM Cloud R11 and revealing some of the new functionality we can look forward to.
Next up was a customer success story from one of Cedar’s recent go-lives – Chubb Insurance. The team spoke about real business benefits from implementing Oracle HCM Recruit (formerly Taleo) and how Cedar and the in-house team worked together to deliver the project within the time-frame. We’re now working with them on a phase 2.
After networking lunch I went to the payroll stream to listen to Oracle’s Andy Spencer introduce the Oracle Payroll Cloud.
The final session I attended was Merrick Hartslief and the benefits of Oracle Social Sourcing (extending Taleo Recruiting functionality).
Before we all headed off to the pub for some well-earned refreshments and sharing of learnings we all got together for a prize-giving, including an iPad, some champagne, etc:
If you missed out and would like to attend next time, please contact email@example.com
HCM Data Loader (HDL) data ready to load into Oracle HCM Cloud is in pipe separated text file format, however most people will create and manipulate these files in MS Excel as it’s the handy swiss-army-knife for data manipulation that almost everyone is familiar with.
The way that we’ve worked is that we create template files containing the sheets and columns corresponding to the fields that the client is using, which the client then populates, then we’ll validate and load into HCM Cloud. Although we use the Data File Validator for HDL as a final pass, most of this validation is first performed in Excel.
Some of this validation is basic (checking for trailing spaces, making sure the provided values are valid in the lookups etc) and some is slightly more complex, eg. looking at consistency across templates. It was whilst doing the latter today that I colleague and I came across a tip that we didn’t previously know.
How VLOOKUP almost works
My normal method of looking up a value in a table elsewhere in Excel is to use the VLOOKUP function. It’s quick, easy and has saved us from countless data issues by spotting problems early. There’s a problem however, which I’ll explain, and provide the solution.
A simplified example is Banks and Branches. First we obtain the list of valid Banks, then we check the list of Bank Branches to make sure that the Bank operating each Branch appears on our list of Banks.
On the left here we have the list of valid Banks in column B, and a text string in column C saying ‘OK’. This is the value that gets returned if the bank matches. (This is a heavily simplified example, the genuine data would have many thousands of branches to check.)
On the right we have the list of Branch owners to lookup against the list on the left.
By using a formula we want to check each of the banks in column E is somewhere in the table on the left. So we use:
=vlookup(<bank to check>, <table of valid banks>, <column to return>, FALSE)
which translates to this for the first cell:
This works a treat, giving the following results:
We can clearly see which branches are run by banks on our list and which are not.
The reason that this sometimes fails is VLOOKUP isn’t case sensitive. Looking up the value LLoyds in the valid banks table would result in a match (despite the second upper case L), however it would obviously fail when we tried to load the data in using HDL.
The method that I now use is a ‘Lookup Exact’. It combines two Excel functions (surprisingly, LOOKUP and EXACT) to give a case sensitive equivalent to VLOOKUP (with the added benefit that the lookup table doesn’t need to be sorted alphabetically).
The formula has the syntax:
=LOOKUP(1,1/EXACT(<table of valid banks>, <bank to check>), <values to return>)
which translates to this for the first cell:
And if I add this to column G we can see that it looks up perfectly against our table, correctly identifying even those in the wrong case (where vlookup in column F fails us).
To give credit where due, I didn’t create this Excel function. It is well explained in this YouTube video: