Taking the self-driving enterprise for a spin - Aera CEO Fred Laluyaux on learning from users, building a category and not being Siebel
- Summary:
- It's been two years of building technology, proving value and learning lessons en route; now it's time to scale up the noise level.
It’s been over a year since I last checked in on the progress of Cognitive Automation start-up Aera. Since then the company has raised $80 million in a fresh round of funding, but has remained largely in stealth mode, despite working with some of the largest ERP users in the world.
All of that could be about to change in 2020, according to CEO Fred Laluyaux:
The last 12 months were 12 months of doing three things, starting with scaling the organization. We really sell to the largest companies in the world and they require muscle, so we had to grow that muscle. We're 350 people now and we're growing at 30 people a month right now, so it's significant. We have worked to really demonstrate the value of the feasibility and the value of Cognitive Automation. I mean we came up with an idea, then we had to go and sell that idea. We had to build the core technology to be able to develop - and I'm using that word properly - develop the first solutions.
Laluyaux has talked previously about what Aera sets out to enable as being “the self-driving enterprise” and he frequently returns to the metaphor of the self-driving car:
Does it actually work? You have to take the car on the road and drive it across the country and see if it's actually delivering the value that you want. That is one of the big things that we've focused on and it's been super interesting because we've learned so much. When you start literally having an engine that is decision automation in real time and at scale, it opens up a whole new world of change management and behavioral management that we knew was there, but we didn't know that no one's ever done it. So you feel like you're opening a new way of doing things. Now it's been done in other areas. It's been done with the banks, it's been done with automated trading. It's been done on the shop floor for a very long time, but it hadn't really been done where we are we operating.
One unanticipated learning that has emerged from this process is that a common question that can be hard to answer inside organizations is that of who owns the numbers? While the business side is held accountable to deliver results based on particular numbers, those numbers are usually not ones that have been created by the business.
Aera can bring about “a shift in ownership” here, argues Laluyaux, with the pendulum moving towards the business, but there are cultural changes that need to be accommodated in many cases as human employees receive recommendations for action generated by automated crawling of ERP systems:
The mindset of people was still, 'Hey, when I get a recommendation, I as a human, I have to do the whole analysis'. Well, what's the point? It's like if every time self driving car technology was giving you a signal to turn left or right, you were doing all the checks yourself and then it's not self-driving.
So it takes a while for people to realize that. Maybe you say. ‘We have a new tool called the Rules Manager that allows you as an end user, as a business analyst, to say that for all recommendations that have an impact of less than whatever threshold - cash, cost, service level - you can either ignore them or you can run them automatically and maybe focus on the ones that have a very big impact or where Aera predicts a big impact. Then you spend more time and insight building the confidence. So all of that is fun because it's never really been done. It's literally like putting a driver in a self driving car and observing their reaction when you say, ' Hands off the wheel'. You've got to learn that So that's what we've done and we've done it with the largest companies in the world, where change is not easy.
Data culture
Some of the numbers that Laluyaux cites from use cases are huge, such as Aera being able to pull out 20,000 recommendations for action in one case. That being so, there must be a risk of ‘drowning in data’ or ‘data hoarding’ taking place among decision-makers on the basis that knowledge is power and declaring ‘ownership’ of data sets for office politics reasons.
Laluyaux says he’s seen that kind of protectionist behavior in the past, but reckons that the situation is rather different this time around:
We know how to crawl billions of rows of data as in ERP systems and we know how to build Cognitive Automation and can demonstrate the value of it literally. Now the fun thing is, we started at one enterprise with one division in one region; now those companies are doing global roll outs around the world with our technology. So the scale is there and we're really moving to how do we fine tune this? It's more a question of learning and fine tuning. So I haven't seen so much politics, right? The one politics is that of, who owns the numbers? But I think for us, at that point it's about explaining how it works.
And that explaining is paying off, he insists, and gaining validation:
Two weeks back, we were at a Gartner event in Denver and everything we've talked about is showing up on their hype cycles and all that stuff. The category is being formed in front of our eyes. I think we've shaped the narrative quite a bit. All this stuff is literally what we've been talking about. You see the Gartner analysts now talking about self-driving supply chain. The narrative is it's getting through. If you work your butt off and if you're lucky and if you have the money to do it, it will take three, four years to establish a category. I think it's shaping nicely right now, but we're still early. We haven't crossed the chasm yet.
That said, if the category you lay claim to having spawned and shaped is getting on the analyst and influencer radars, then you’re also likely to be on those of the established ERP vendors. At that point, the question from history is simple - how do you make sure you’re Salesforce and not Siebel? In other words, how do you own leadership of a category you pioneer and not be usurped by challengers, as Oracle did to Cullinet in the database market, for example.
Laluyaux concedes the point, observing that there are two goals:
You have to establish the category and then you need to be the leader. How do you do it? We've devised a plan to do it and we're trying feverishly to execute it and then I'll tell you at the end if we did it or not. We are boxing with the big guys here. We know they already know that they should be in it.
So, it’s about, ‘Good for you. You came up with it. Now how are you going to win the race?’. What we've done is we've taken our time to establish. We brought in $50 million just to start and I knew it was two years work. Of course it's not two, it's three, but that's okay. If you're not 50% delusional as an entrepreneur, you’ll never grow, it's too painful. I invest in a bunch of companies and I always say, ’It’s going to take twice as much money and twice as much time, if you're lucky, if the gods of entrepreneurship are with you’. .
What I would be worried about is if we had fumbled on this and found that we're not ready. We've really designed the company from its board, from its team, from the get go, to actually be able to achieve that. It's not simple as a start-up to go and sell multi-million dollar projects with the Fortune 500. It's incredibly hard.
Countering that is about tapping into experience, he argues:
We brought a team of folks who've been together for a very long time and know how to do that. So we've already scaled organizations. We've all worked for SAP, we all built Anaplan. We've all brought that DNA. I don't think this company would be possible if we hadn't done everything that we've done before. We would break under the pressure if we hadn't been there before.
You know that you're going to start running a marathon, you're going to go through pain, but you know that you're also going to potentially finish the race. So, we've brought another 80 million bucks this year. It plays very long. So we have the time, we have the money; It costs a lot of money to build a team of 300 or 400 people, to basically keep the business growing. We took the time to really first demonstrate the value and build a muscle with a few accounts and be successful before going to scale.
He adds:
I think a few years back I would have tried to scale sooner and I would have failed. This is very, very simple physics where the higher you go in a company, when you start selling to the C level, they all know each other, they all talk to each other. You cannot afford failure because one mistake could cost you a lot. So we took our time. We have a very deliberate strategy. We have a big milestone coming up next year when we think that will be the time for us to really come out and scale up to more noise. We've been very quiet and focused on building the technology and demonstrating its worth.
The timing is ripe for a firm like Aera, he concludes:
Today we can crawl 47 different ERPs for one of the largest Top 10 market cap companies in the world - 47 ERPs, real time, You can see the value that we deliver every day. The system computes that value. It's real. There's no fluff anymore.
My take
There’s no doubt that Aera has put in the work to build out its technology offering over the past couple of years and there are some big name use cases, such as Merck and Johnson & Johnson, on an increasingly long list. It’s encouraging now to hear Laluyaux talking in terms of ‘making more noise’ in 2020 as it’s time to cement ownership of this category before the ERP establishment circle their wagons. How it executes on that challenge will determine the outcome of the next chapter of the firm's growth.