As my colleague Dennis Howlett notes in a previous post, Alteryx's mantra is 'data scientists not required'. And although many will argue that this doesn't align with what others are pushing for in the 'big data' market, when you think about it, it makes a lot of sense. Giving line of business users control of their data and getting them to use that data to make better decisions, easily, can deliver some quick wins. As Dennis notes:
Customers routinely report that the workflow elements of Alteryx condense data modeling time dramatically and the time to outcomes fall dramatically. Ingersoll Rand for example was mentioned as a company reducing the time to review text fields down by 5-6 hours per report. But it was in the breadth of use cases beyond consumer data analysis where Alteryx impresses.
Alteryx's President and COO George Mathew, who used to work for SAP Business Objects, said that the company's success has been driven by this shift in the market.
As I've seen this market evolve, it's been quite remarkable because I came from what would be that older school BI thinking. I really started to notice that the front end of that BI market is actually shifting quite remarkably away from more authored solutions like Business Objects, Cognos and Hyperion into more self-service, self-reliant visual analysis dashboard capabilities.
What I asked myself was a very simple question back then: why isn't the underlying data preparation, data blending, analytical modelling also going towards more of a self-service reliant manner, enabling the same line of business users? Why is it just about the visual dashboarding and analysis experience? Really that's where Alteryx has managed to create a really stunning amount of growth.
Escaping ExcelMathew explains that Alteryx is popular in environments where customers are attempting to blend and analyse a variety of data sources, with customer transaction data being an obvious use case. He added that 60% of the company's user base are day in and day out dealing with a least five different sources of data and are using the platform to blend these and create analytical models. He says:
“Having to deal with a variety of shapes and a variety of sources of data at the same time, that's when Alteryx becomes very, very valuable for those line of business users.”
Mathew says that whilst customer transaction data has been a popular use case to date, the word is spreading internally at these organisations and is being used across other functions such as supply chain and HR data. He believes that Alteryx should be being used whenever an Excel VLOOKUP is the first choice.
Mathew doesn't think that Excel should be being used for complex data modelling. He says:
If you ever have to use a VLOOKUP in Excel, that's a first awesome candidate for Alteryx, because VLOOKUP is usually the first form of data blend. I don't think there's anything fundamentally wrong with a spreadsheet as a data scratchpad, it's what it is. Most spreadsheets are awesome data scratchpads.
The problem is that spreadsheets have become data sources, they have become the place to do very complex analytical models. In those situations it becomes a very unwieldy scenario. If you're an analyst that's created a sophisticated model with a spreadsheet today, the chances that the next analyst would want to pick up that model and unpack what's inside the spreadsheet are actually pretty slim. The analyst that is picking up that work would rather just rebuild the model.
We literally start with a customer in a line of business situation where one user starts to use Alteryx for a more sophisticated blending situation than two Excel sources of data being brought together – once they have been successful they tell ten friends inside the organisation and a few friends outside the organisation. A lot of our spread in the market today is really driven by grassroots, ground up, viral adoption of a product.
According to Mathew, evidence of the rate at which people are embracing Alteryx can be seen in the speed of adoption. For example, 92% of customers are building their first analytical processes within five days of downloading Alteryx; 45% are doing it within the first four hours; 22% are doing it in the first hour.
However, Mathew admits that although adoption and spreading the 'Alteryx word' has had traction over thepast couple of years, there are still challenges in getting enterprises to fully become 'data decision makers' – where line of business users feel empowered to take control of their data and use it to regularly make decisions. Once again, it's culture that's the problem. And getting around that can be tough. He says:
There are a few things that continue to be a challenge, but interestingly none of them are technology challenges. I think the biggest challenges are organisational and cultural and about how people think about analytics today. I think we still have a real issue in our market in terms of having a culture of data in our organisations. Invariably the biggest barriers to self service and self reliance are people asking: 'I'm going to be enabling someone downstream to work with data in a way that I don't control directly?'. 'What does that mean?' 'Is that a loss of control?' 'Is that a security issue?'
In reality I think most modern data workers today would rather have the ability to have data in their hands to be able to shape, to understand and model. And not be prevented from that because of whatever the underpinnings of the culture is. I think that's fundamentally the biggest challenge.
However, he adds that this is getting easier because organisations are starting to recognise that there are serious incentives to becoming data driven.
I think some of it is a lot more straight forward today than it was before, because people are seeing the value in being able to put data in the hands of the users. The effectiveness of driving faster analytical decisions. I think it's very well understood in the market that the organisations that use analytics to drive decision making in their respective organisations end up from an industry standpoint doing two to three times better than the folks that don't.
There is a industry-level mandate in place to take advantage of analytics today in a much more meaningful way. The organisations have realised that they have to make a mental shift to be able to get towards better data and decision making, which can drive them in a more financially positive way forward.
FocusAs Alteryx approaches 1,000 customers in the next few weeks, it is also preparing for a big release, which is currently in beta. As well as including support for emerging data platforms, such as Spark and Hadoop, the new release is going to focus on building on an engaging UI and increasing collaboration features, to support the 'grassroots' nature of Alteryx's success.
First and foremost, we are continuing to invest around making the experience around data analytics smarter. How we think about data analytics smarter is much more around the UI/UX of how people can take advantage of analytical tools on a day in day out basis. How do we make sure that within the first five minutes of the download of the product it is as meaningful and impactful as possible?
We actually ended up hiring video game designers inside of product development and management teams, largely to help us to evolve into exactly what a great video game does – how do you level up? How do you become more attracted in that first few minutes experience where you are stuck to wanting to play the game even further? This is actually quite useful when you think about enterprise software today.
He said that this has already proved to be successful, given that the beta release has already had five times more subscriptions than any previous release of the product. Mathew adds that this investment takes up 20% of all the company's development and research efforts.
The second area is how we feel that power users really need to continue to have the ability to share and collaborate with each other. The server within Alteryx is really about having better version control, better sharing, better collaboration of processes that are built by one analyst to be able to share it with the next 20 analysts.
However, with Alteryx adding 200 customers a quarter across a number of geographies, the company's biggest challenge, according to Mathew, is managing that growth effectively. He said:
There is a fundamental challenge when it comes to a high growth situation that you have to keep motoring on as things are changing all around you. You are talking literally thousands of users coming on board any specific quarter. That velocity is quite significant because you are now starting to come across things where you have to make sure they not only have the great experience for the initial download of the product, but also they are supported as continue to become further wanderers and explorers.
You want to make sure that your support function is available around the world 24/7, you have to make sure that the product is internationalised in a variety of countries and languages where historically it had an english-only underpinning. These are very much 'first world' challenges, but are still work that you have to put in. You are doubling the R&D team, you are doubling the salesforce, you are doubling the marketing expenditure.
My concerns around Alteryx echo Dennis' in his previous piece. It will be interesting to see if Alteryx can stand on its own two feet and be a successful platform, without having to cosy up so closely to Tableau. But I think as it starts supporting for more platforms and integrating with a wider array of systems, Alteryx will find a level of independence in the market. Equally, whilst self-service data analytics will appeal to some, others will persist with more complicated scenarios that involve huge data lakes and top data scientists. What works for some, won't work for others.
But I do like the appeal of Alteryx's pitch. Not having to find those valuable data scientist skills and quickly being able to give people control of their data, in order to support better decision making, is something that many companies will quickly sign up to. And less face it, the fewer Excel spreadsheets the better.