The Evolving Role of Architects

In the last couple of weeks I’ve had the great opportunity to spend time with IT architects of various sorts both inside and outside of the insurance industry. The discussions have been illuminating and offer different visions and futures both for technology that supports insurers and for the future of the architecture function in insurers.

One of the main events that allowed for this conversation was a round table held in London with architects from insurers. The main topics were the relevance of microservices style architectures to insurance, the role of the architects in AI and InsurTech and the future role of architects at insurers. Another event that offered an interesting contrast was the inaugural London Software Architecture Conference which I'll call SACon below (Twitter feed).

Microservices

I won't fully define microservices here but briefly it’s an approach to delivering software where each service is built as it’s own application which can be scaled independently from other services.

Microservices as a way of delivering software was the default approach at the SACon. There were sessions where architects sharing stories about why sometimes you had to work with a monolith or even making the case for not having the services in discrete applications. Meanwhile at the round table the monolith was the default still with the case being made for microservices in some parts of the architecture.

There are use cases where microservices make a great deal of sense, particularly in already distributed systems where a great deal of data is being streamed between applications. Here the infrastructure of microservices and the libraries supporting the reactive manifesto such as Hysterix and Rx* (e.g. RxJava) and indeed one insurer related their use of microservices to support IoT. Others discussed using this style of approach and the tooling surrounding these architectures to launch new products and increase change throughput but in all cases these were far from replacing the core architecture.

For now microservices is not the default for insurer software but is certainly a tool in the box. An observation or two from SACon from those looking to adopt: First it doesn’t solve the question of how big a service or a component is, something architects need to discuss and refine and; Second, microservices needs a great deal of automation to make work, a topic covered in our DevOps report to be published shortly.

Architects and AI

I have a background with training and experience both in computer science, AI and machine learning. One thing that I noticed going to the analytics conferences where AI is discussed is the absence of IT representation – plenty of actuaries, MI/BI folks, marketing folks – was this a place for architects?

Most insurers present at the round table had activity within the organisation for AI. For the most part only data architects are involved in this discussion – AI being distinct from business and applications architecture for now. It’s my opinion that AI components will form part of the wider applications architecture in the future, with AI components being as common place as programmed ones.

Architects and InsurTech

Here is an area where architects can more immediately contribute in a meaningful way both in reviewing opportunities and unique capabilities from InsurTech firms and in discussing integration where acquisition rather than investment is the goal.

The challenge here of course is the age old challenge for architects – to have a seat in the discussion the architect function needs to demonstrate the value it can bring and it’s internal expertise.

Finally, one amusing discussion I had was with a few architects from startups. As I discussed legacy systems they also related seeing legacy systems in their organisations – albeit the legacy systems were 2 or 4 years old rather than 20 or 40 years old. The intriguing thing here was the reasons for them becoming legacy were the same as insurers – availability of skills, supportability and responsiveness to changing demands. It may hearten architects at insurers that start ups aren’t immune to legacy issues!

 

 

One last look back at Google Compare

It’s old news by now that Google is shutting down Compare, its financial services and insurance comparison site. It wasn’t open long – less than a year. When Compare was first announced, the industry reacted with warnings that this was a major disrupter in insurance distribution. With the massive audience that Google has, the industry expected that Google was going to swoop down and capture the online insurance market – which by the way is pretty big – typically 75% of prospects research online and 20-25% of all new auto policies are purchased on line according to those who track this type of metric.   So what happened? Well, the fundamental idea of capturing the online market is a sound idea. And Google was pretty smart at avoiding all the hard technical costs of building out the aggregator engine by partnering with those who had already done the hard work – like Compare.com, Coverhound and Bolt.   But the business model of an online aggregator is hard. There are three models – online agents – who earn full commissions. That wasn’t really Google’s deal. They weren’t interested in any of the after service or ongoing relationships. A traffic generator – sending a potential lead to another site and being paid for the eyeballs. Well, that’s not very lucrative either – and frankly, Google can make money through their own advertising and search capabilities. Spending the money to build an online quoting front end only adds cost to something they already do quite well, thank you.   So why would Google have invested the money in an online quoting front end? To take advantage of a lead model. With a lead model, the aggregator collects data, processes a request for quote and sends a highly qualified lead to be fulfilled. The price per lead is significantly higher than the price for traffic. But there’s a fundamental challenge with this model. For the lead to be valuable to a carrier, the lead has to actually purchase insurance. And because a lead is sold to multiple carriers, the acquisition costs rise for a carrier.   Let’s say a lead is sold for $5 to ten carriers. The aggregator makes $50 for that lead. But only one carrier actually writes the lead. If ten leads are sold, and each carrier writes one, the aggregator makes $500 but the carrier has spent $50 for that lead. Play out a competitive situation where the leads aren’t equally distributed, and you can see that the acquisition costs can rapidly rise. If I only get one lead out of twenty, I’ve spent $100 for that lead. If I only get one lead out of $30 I’ve now spent $150 for that lead – which now is pretty close to what I’d probably be paying an independent agent. And what if the customer NEVER buys – and simply goes in looking for prices so they have a comparison to an off line model? The numbers rise rapidly. Remember those numbers above – 75% shop on line and 25% purchase on line. That means that only one in three leads actually results in a sale. Assuming leads are distributed evenly, an aggregator will distribute 165 leads before I close one. That brings this $5 lead fee up to $82.50 –, which is pretty expensive. The way to make those economics work is to increase the conversion rate so that more of the leads a carrier purchases actually ends up buying a policy.   So while carriers are very interested in participating in the online marketplace, they really want to work with those aggregators who are successful at converting traffic to leads that will convert to policyholders. The online agent model is attractive as the carrier doesn’t pay until the policy is written. The traffic model is similar to online advertising, so that works as well. But the success of a lead model is a combination of the price of the lead and the likelihood of closing that lead – which is dependent on the number of carriers the lead is sold to and the propensity to buy.   So here’s where Google lost an opportunity with Compare. They thought they could convert relatively low paying traffic into high paying leads simply by putting a quoting front end on and didn’t think through what they could have done to improve the conversion rates. With their analytical power, Google could have created a truly disruptive experience by providing consumers with a powerful recommendation engine. Google is a master at finding out information about individuals from social media and other publicly available data. They could have created an algorithm that used the information about the lead to tailor and target recommendations.   Personal auto isn’t that hard. If we were talking about commercial, it’s a much harder set of algorithms. But honestly, it’s not that hard to create something that tells a customer that given their location, the value of their home, the type of vehicle and their driving record, 64% of people like you choose this limit/deductible/additional coverage etc. And getting a personalized recommendation drives conversion. When people trust that the advice is good, they’re willing to buy. We’ve seen many examples of how inserting advice and recommendations into the quoting process drives conversion.   When I personally go to get an online quote – it’s part of my job – I enter information that shows I own a home in California and I drive a luxury car. So why oh why do the aggregator sites today recommend minimum limits coverage to me? My car is worth more than that. Today, trusting the advice from an aggregator site is dicey. And that is why policyholders continue to rely on the advice of an agent. Does this mean the role of aggregators is dead? No.   But Google missed a major opportunity to truly disrupt by providing a powerful recommendation engine that could use their ability to easily find information about individuals and combine it with their powerful analytical abilities. They ended up creating just the same thing we had back in the 90’s. Kudos to them for killing it quickly – but they missed an opportunity to use their capabilities to make the model work.  

Pushing beyond apps

It struck me while I was driving this morning: First-gen mobile apps are fine, but virtually everyone is missing high-volume opportunities to engage with their customers. Allow me to back up a step. I was stuck in traffic. Not surprisingly, that gave me some time to ponder my driving experience. I found myself thinking: Why can’t I give my car’s navigation system deep personalizations to help it think the way I do? And how do I get around its singular focus on getting from Point A to Point B? I explored the system while at a red light. It had jammed me onto yet another “Fastest Route,” disguised as a parking lot. My tweaks to the system didn’t seem to help. I decided what I’d really like is a Creativity slider so I could tell my nav how far out there to be in determining my route. Suburban side streets, public transportation, going north to eventually head south, and even well-connected parking lots are all nominally on the table when I’m at the helm. So why can’t I tell my nav to think like me? I’d also like a more personal, periodic verbal update on my likely arrival time, which over the course of my trip this morning went from 38 minutes to almost twice that due to traffic. The time element is important, of course. But maybe my nav system should sense when I’m agitated (a combination of wearables and telematics would be a strong indicator) and do something to keep me from going off the deep end. Jokes? Soothing music? Directions to highly-rated nearby bakeries? Words of serenity? More configurability is required, obviously, or some really clever automated customization. Then an even more radical thought struck. Why couldn’t my nav help me navigate not only my trip but my morning as well? “Mr. Weber, you will be in heavy traffic for the next 20 minutes. Shall I read through your unopened emails for you while you wait?” Or, “Your calendar indicates that you have an appointment before your anticipated arrival time. Shall I email the participants to let them know you’re running late?” Or (perhaps if I’m not that agitated), “While you have a few minutes would you like to check your bank balances, or talk to someone about your auto insurance renewal which is due in 10 days?” What I’m describing here is a level of engagement between me and my mobile devices which is difficult to foster, for both technical and psychological reasons. And it doesn’t work if a nav system is simply a nav system that doesn’t have contextual information about the user. But imagine the benefits if the navigation company, a financial institution, and other consumer-focused firms thought through the consumer experience more holistically. By sensibly injecting themselves into consumers’ daily routines—even when those routines are stressful—companies will have a powerful connection to their customers that will be almost impossible to dislodge. Firms like Google have started down this path, but financial institutions need to push their way into the conversation as well.

Google Glass is dead–long live Google Glass!

On January 15, Google announced that Google Glass was not ready for prime time. Google is withdrawing it from general availability, but says it will relaunch at some unspecified future date. With the benefit of unaided 20-20 hindsight, Google Glass was a consumer product designed by engineers to appeal to, well, other engineers. Consumers who bought the $1500 device, started to be called a number of things, but “cool” was not one of them. The form/function of Google Glass (a hands-free, heads-up display, sourcing various kinds of data and information, with a video recording capability) suggests several business uses, for example, in insurance: field loss estimators and loss control engineers. And there are of course many other professions with similar needs (e.g. service and repair technicians). The form/function of Google Glass will live on, probably in several diverging incarnations. But whether any of these descendants will be recognizably a “Google Glass” or whether their wearers/users will ever be cool; remains to be seen.

Does Google Know Your Religion?

An industry contact recently told me that her phone popped up the following “creepy” message one Sunday morning: “9 minutes’ drive time to St. XXXX Church.” This, of course, was a predictable result of Google Now keeping track of where and when she regularly went, and nobly trying to help her get to her regular 8:30 a.m. Mass on time. What makes it creepy is that most people put religion (and politics? and sex?) on a mental “Off Limits” list. Deeply personal issues like these are risky fodder at cocktail parties, and equally risky subjects for automated, data-driven insights. An application like Google Now doesn’t understand the issue unless its human coders are prescient enough to realize there are some connections we humans simply don’t want our devices to make on our behalf. But now the string of unintended consequences has begun. My contact told her story in a room full of people. The consensus reaction around the table was, “Google keeps track of where you drive, and even figures out what’s there? That’s over the top.” What else, we wondered, would the app notice about us and dispassionately reflect back via pop-up message? Some bland examples: “You’re almost out of gin, and there’s a liquor store nearby…” “Your wife won’t be home for another hour, and Melrose Place is on channel 7 right now…” “You’ve been getting a lot of emails from XXXXX—perhaps you should ask him/her out?” The conversation took us quickly from ambivalence to unease. Others in that group probably mentioned their unease to their spouses and colleagues and friends. As I write, an ever-widening circle of people is developing reasons to be suspicious and uncomfortable about Google Now—which, by the way, is a perfectly excellent and useful app about 99 percent of the time! It’s a grand example of the problem we’ll face as we try to harness the power of big data. Years ago, when desktop publishing first became a technical reality for business users, a friend of mine who was a professional designer put a sign on her wall: “Power Ability.” She was telling people that just because they could publish their own office newsletters, crammed with cutsie clip-art, didn’t mean that they should. Someone might want to give Google the same sort of advice, especially when church locations are involved.

Which U. S. P&C insurance company will be the first to use a social network as a platform to transact insurance?

Part of the hype involved with the Facebook IPO this week is the vision of its founder to establish it as a platform for people to use when interacting on the internet. To me, this means not simply linking from the S.N. to another site, but to actually complete commercial transactions on the social network itself. For insurance, this would have a minimum benefit of avoiding tedious data entry of demographic information that the network already knows about you, age, sex, address, etc.

If you don’t think transacting insurance on a site other than that of the writing company will happen, you can stop reading now — you won’t be interested in the rest of this post. If, however, you are game to consider that a company will try this, read on and get ready to post a comment and make your pov known.

What would be the characteristics of the first U.S. P&C insurance company to use a social site as a platform for its business? Here are some considerations:

Product – Will be one of the commoditized, less differentiated products such as personal automobile, motorcycle, or dwelling: In these lines of business, price and service separate the offerings. Being able to ease the process of obtaining and maintaining insurance will be a driver.

Segment – Likely youthful: The companies first attracted to a S.N. platform will be those with prospects who are most comfortable with the environment, aggravated by traditional insurance distribution and less concerned about personal privacy.

Channel – Direct: As this new way of doing business will involve first-mover risks, it is likely that companies with intermediated distribution will pass on using the platform in the beginning. Upsetting their existing agents with a direct social network approach will be too much to bear.

Culture – Innovative: There will need to be some match between the value system of the insurer and that of the social network company. Stepping into this unknown territory will require that both parties are comfortable with their partner. Thus, insurers which have a reputation for being more innovative will be more likely to reach acceptable terms with the platform provider. Since this will probably be either Google or Facebook, the insurer with the vibe closest to “do no harm” or “hacking” will likely prevail.

Size — Not necessarily the largest: I do not think that the first S.N. insurer will need to be a billion+ organization, but I am sure that the network provider will demand some fairly steep rent and this will restrict the number of insurers able to pay the freight.

Technology – Probably not a barrier: The investment of many insurers in core system renewal should position most of them to take advantage of the open standards that a S.N. is built on.

Regulator view – The wildcard: The most likely insurer, in my view, is one that has a good reputation with the regulators and a decent reputation at addressing their concerns in past market conduct reviews and inquiries. As the response of regulation to this new way of doing business is such an unknown, the company willing to take this step will be confident in its ability to respond to its overseers.

Given these parameters, does anyone come to your mind? In days past, I would have identified the company with the pink-haired lady mascot as a leading possibility, but its recent purchase takes them off the table. It also could be one of the specialty auto insurers targeting the youthful driver market as well.

Post a reply, or email me at mfitzgerald@celent.com, or contact me on LinkedIn or Twitter and let’s knock this around a bit.

The Implications of Going Driverless

Google recently received a patent for a driverless car that can handle the challenges of the open road. (Story: Google Granted Patent for Driverless Car Technology) I love this story, and the implications.

The NTSB’s proposed ban on texting will require a rewrite. Unless my autopiloted car and my smartphone are sharing computing power or connectivity, I’m assuming my texts sent while underway won’t impact safety. Of course, by the time this system is operational, texting will most likely have become passé.

The field of play for telematics is about to get more complicated. Sure, you can track where my car went. But was I at the helm, or was my car driving itself? And what if I get out of the car to let it park itself? Does that still count as me driving, for insurance purposes?

If my car can drive itself, under what circumstances will I even decide to go with it? For example, for many daily errands (e.g., picking up the dry cleaning, grocery shopping, taking the kids to football practice) the main value-adds I bring are navigation and execution of the route. Take those away, and I might choose to do something else.

We’re one step closer to answering one of life’s Big Questions. Namely: Are men or women better drivers? The answer may turn out to be neither, assuming that Google gets the product right. With R2D2 as our chauffeur, my wife and I will have to find something else to debate. Related question: Will my autopilot car stop and ask for directions when it gets lost, or will it just drive around hopelessly? For tradition’s sake, I propose to make that a user-configurable feature. I’d hate to have my car out-perform me in such an obvious manner.

Hey Facebook and Google, why I'm Liking it and looking forward to +1ing it

Perhaps it’s because I like technology, because I was born into an age of unprecedented technological advancement, because I’m curious or simply because I’m too lazy to keep in touch with folk but I have to say I love social networks. For me they’re a tool that allow me to keep folks up to date and to keep track of what friends and acquaintances are up to. The whole thing was brought home to me recently when I was tinkering with ancestry.com and talking to my family about my extended family. On a whim I had a quick look in Facebook for my Mum’s cousin and found her, alive and well in Australia. A few messages on Facebook later and I discovered she’d had a few children and there were grandkids dotted around Australia and the US. The real value of Facebook came home to me when I was able to sit down with my 3 year old boy and show him the pictures distant family members had shared and show him where they lived. All this and I can keep them up to date without doing more than i do today – got to love Facebook for that fire and forget, status update to everyone feature. So why use the Facebook Like button? I think about the Facebook Like button in a similar way to Amazon’s suggest feature. Every now and again I go on Amazon and tell it things I would like to get, things I’ve purchased and even rate some of the things. This investment pays dividends in relevant suggestions from Amazon on books and other items genuinely of interest to me. Facebook Like allows me to share likes with my friends and allows Facebook to suggest things I might like, so recently it suggested a bunch of my friends like Terry Pratchett’s Facebook page (I’m a fan of the UK’s most prolific author) and I happily discovered a new book is due out shortly and some of the other activities Terry is up to. For me Tweeting, Updating, Liking, Following, Friending – it’s all about filtering the wealth of information out there to find the bits I’m interested in quicker. I make investments by creating content and sharing it – like this blog post, and for my very small effort I am typically rewarded ten fold. This appeals to the lazy efficient part of me. This is how many (though not all) of my friends are using social networks today. There are issues with all this though. Different groups of people are interested in different things – I know colleagues, ex-colleagues, friends, family, folks who live in my village, people from university, school – people interested in games, technology, mobile phones, insurance technology and wierdoes as Craig Weber described them. Oddly enough I don’t know anyone who is all of these things yet these networks treat them all the same. To some degree using Facebook for some things, Linkedin for others, twitter, skype, etc. kind of works but this presentation on a version 2 of a social network really spoke to me. Google have just announced Google+ – something I’m looking forward to trying out because of a few key features:
  • circles – group friends in different ways so you can share content but only send it to those who’ll be interested. I’m looking forward to creating the wierdoes circle
  • sparks – a feature that claims to go and find content for your interests and pull them together – awesome – why search when the relevant web content can come to me?
  • hangouts – 10-way video chat! Need I say more.
Google+ is in it’s infancy and the literature lacks some of the business focus but I’m sure Google+ will find it’s way into Google Apps for Enterprise in due course. What should an insurer take away from this? The hidden subtlety here, and the reason Google has been forced to respond to Facebook with Google+ lies in sparks, or Facebook like suggestions or Amazon suggest. Internet users are moving away from searching. Brands focussing solely on search engine optimisation will lose out, as will company’s focussed on search. In the future systems will suggest content, reviews, products, brands and insurers to customers based on their behaviour and social circles. Whilst today’s drive to get Likes and reviews seems shallow and immature, it points to a fundamental shift in the role of the Internet in driving the acquisition of new business. It’s hard to say where all this will fall but a few things are clear:
  1. Whether we like them or not facebook and social networks in various guises are here to stay
  2. Google has re-entered the social network space, even if it’s not successful (like orkut?) the new features will change other social networks
  3. Social networks and they’re features are changing the way we interact with the world, each other and with insurers. Like Google, the insurance industry will be forced to respond
For more look at our coverage of social, read why Craig Weber isn’t using Facebook and look out for more on the value of Facebook pages and the social internet for the insurance industry in future posts and research. I’m off to tweet, like, post and +1 this blog post.

Social Media so hot, Ben & Jerry's email marketing melts

The story that Ben & Jerry’s are dropping email marketing in favour of social media hit something of a sweet spot with me. Not only do I not like trawling the ever increasing mass of emails each day but I also have a keen interest in how the Internet is evolving and a highly developed sweet tooth. The story is quite interesting as it tracks some of the trends Celent observed in our Digital Marketing in Insurance report. For many insurers email marketing and communication is the primary digital method of reaching consumers, however most insurers saw social networks and social media as becoming an increasingly important channel to market. Perhaps Ben & Jerry’s move is both a little early and a sign of things to come. This also comes at a time when rumours abound regarding yet another social network set to come from Google, possibly to be called Google Me. Indeed there are stories that disenfranchised companies working with Facebook may already be signed up to work with Google Me. A slide show published by a user experience researcher at Google offers some insights into the key issues with current social networks, how the new social network will look and the features it will offer businesses. For insurers looking at social networks as a medium to customers or looking at how they can expand their use of social networks Celent’s report on the subject may well be of interest. Addendum: Also making waves in social media is the old spice campaign on YouTube. Effective use of this style of campaign is discussed in our report but it is particularly well executed here.

Can Google Save Us From Bad Interfaces?

As often happens, a recent conversation with an insurance CIO produced an “ah ha!” moment for me. Why is it that we, as an industry, are so slow to recognize technology user behaviors that have become commonplace? If everybody Googles, why don’t insurance applications?

The topic of discussion was the difficulties that technology execs face in supporting the polarized needs of users. For example, some agents still prefer to take new applications via their trusty yellow legal pads and ballpoint pens, while others take a laptop-centric view of the world and prefer electronic applications. Those two streams may converge at some point (probably on the desk of a field office or underwriting assistant), but there are critical implications for processes and technology.

Even within the subset of agents that is technologically savvy, there are differences. For example, most agents are comfortable generating an illustration using traditional desktop software, which presents a series of data entry fields and spits out compliant presentation materials in about 10 minutes. But what about the agent who is a Google power user, who thinks in terms of a command-line interface to do things more quickly and more directly? Could a Google-based comparative illustration application work?

On a goof, I tried entering the following in Google: “male 44 year old non-smoker 20 year YRT.” What the heck, I thought, go with the I’m Feeling Lucky button. (For the uninitiated, this slightly irreverent option in Google takes you directly to the top search hit based on the other terms entered, rather than presenting you with a list of all hits.)

I’m Feeling Lucky produced no result. Shocking! Apparently, there are no carriers leveraging a pure online illustration package with a Google interface. A regular Google search wasn’t much better—it turned up lots of articles on mortality and insurance, but not what my hypothetical agent was looking for.

The lesson for me is that if the insurance industry truly designed technology to meet user needs and expectations, this application or something similar would exist. The technology is proven, and commercially available.

Is insurance more complex than searching for a JPG of a common house wren? No question. Are there difficult issues with designing a business model around this approach? Absolutely! But technology exists to solve problems, and sometimes to enable radical shifts in how businesses and customers interact. We could avoid a lot of pain and effort by leveraging existing user behaviors, rather than forcing users down the paths that are most convenient for us.