Much ado about upload, download
There’s been a lot of hype and misleading information getting tossed around in the trade press on the electronic interface issues. Agents are constantly being told to “automate or die,” but little attention is being given to the underlying costs, management concerns, and practicalities.
To quote a subtitle on a recent article on single entry multi company interface (SEMCI), “Agents who’ve already launched into interface space are rocketing towards increased profits.” Mighty powerful words, and maybe true in some cases.
I’m not questioning whether company/agency interface is a good idea. My concern is that many ideas are being published as matter-of-fact truths, when the actual truth is far less clear. And I’m afraid that some people are rushing into something with only a fuzzy or partial understanding of what is really going on.
Let’s throw some light on the subject and see how the average agent fits in.
What is interface?
Agents have had company terminals in their offices for years. Is this an interface? Of course it is, but is this the interface everyone is talking about?
The activity that’s getting so much attention presently is interfacing a company’s computer data to an agency management system’s computer data. So let’s use that as our working definition of interface.
Lesson Number 1: Be sure you know what is being talked about before you try to apply it to your situation.
Next, we need to put some things into perspective, in terms of agency needs. On one end of the spectrum, we have mega-agencies like Alexander & Alexander, Fred S. James, etc., that have tremendous quantities of data to manage. It doesn’t take a detailed cost analysis to figure out that these agencies can benefit financially from electronic interface. And, not surprisingly, this is where much of the early interface work was done.
While these large agencies account for a large number of premium dollars, they account for only a small percentage of the total interface partners needed to say we really have a global interface standard.
The more typical agency might be KAB Insurance Agency, a $2 million personal lines agency in Alpharetta, Georgia. KCB has about 2,400 policies in force, averaging about $800 each in annual premium. (Please feel free to play with these numbers yourself, making them match your situation.)
Suppose KAB’s policies are fairly evenly split between Company A and Company B, placing 1,200 policies with each company. With things the way they are today, odds are that, at best, only Company A will offer interface.
So for Company A, KAB has 1,200 policies. Assuming some activity on each policy once per month (do you have activity on every policy once per month?), there are 100 transactions per month, or about five per day. Really, now, how long will it take a CSR to process these five transactions even without the benefit of upload/download? Is automating these five transactions going to dramatically increase the agency’s profits?
And finally, how much can the agency spend to automate five transactions per day? Can KAB afford to spend $40,000 or more on a full agency management system for the express purpose of interfacing these five transactions?
I don’t see how agency profits are going to skyrocket at these levels.
Lesson Number 2: Do your own arithmetic. Does it work for you?
What do the statistics mean?
I’ve heard there may be 60,000 independent agencies in the U.S., and maybe 3,000 P&C companies. Let’s look at some published statistics in that light. (Again, substitute your own numbers if you want to.)
First is the concept of “download pairs” or “upload pairs.” This is simply the number of agency/company combinations. For example, if each of the 60,000 agencies does business with 2 companies, then there are a total of 120,000 potential “pairs” out there. If each does business with 3 companies, then there are 180,000 total pairs.
Quoting from ACORD’s 1994 Automation and Interface Survey Report, there are 2,881 download partnerships. As published in the May 1994 Rough Notes (page 15), Applied Systems has 1,741 download pairs, involving 43 companies. These numbers seem to indicate that Applied has well over half of the download pairs in existence. Applied is certainly among the leaders in developing interface capabilities, and they deserve a lot of respect for the work they’ve done.
I don’t want to minimize the accomplishments thus far, but 1,741 pairs out of 180,000 is a drop in the bucket (less than 1%). Forty-three companies out of 3,000 is still only slightly over 1%.
Allowing for all the rest of the work done by other companies and other vendors, their efforts may have penetrated only 2% of the marketplace. (Don’t let claims measuring the total premium volume confuse you–that’s just the way the big guys impress each other, and it further goes to prove that the needs of the mega-agencies may be different from the average agent.)
To read the trade press, you might think you’re about the last agency out there that isn’t downloading yet, and if you don’t start soon, then you won’t be in business much longer. And you should be pressuring your company to interface now, while you still can. Well, welcome to the other 98%.
Lesson Number 3: Don’t let anyone tell you what the numbers mean. Look at the specifics and decide for yourself.
Who’s it for?
Upload and download serve two very different purposes, although they are at times related.
Upload is undoubtedly for the company. The agency does the data input for them, which reduces the company cost and probably increases accuracy. The agent is going to have to type the data into a rating system, or maybe an ACORD form, so maybe it really doesn’t matter where the agent types it, as long as it has to be done only once.
Download is mainly for the agency, and is a by-product of the work done in the company’s computers. One company representative recently said in public that he didn’t care about download at all, except that it was needed to provide a good starting point for upload (which he cares about a lot).
The system vendors don’t really care about either upload or download–they just want to be sure the competition doesn’t get far enough out in front that they lose business.
Which brings us to a crucial point: From conversations I’ve had with agents, insurance company and agency system vendor representatives, no one wants to pay for interface. Companies using mainframe technology have to spend enormous sums of money just to put in interface capabilities in their own internal systems. System vendors likewise have a major investment to ready their systems to accept the download. So why is everyone standing around, looking at the agent and smiling a lot?
If someone is telling you how valuable this is, and why you should plunk down large sums of money for it, ask them why they don’t pay for it instead. Where I come from, if no one wants to pay for something, somebody starts asking “What’s the point?”
Lesson Number 4: If you think interface is important, be sure you understand what you give and what you get.
Let’s not overlook the much-talked-about ACORD standards.
ACORD ventured far out into the interface arena several years ago when it chaired the project to develop the ACORD Level 3 interface standard, known as AL3.
This “standard” is simply an agreement between companies and system vendors that they will speak in common terms when electronically communicating with each other. Developing such a standard is certainly desirable and even necessary to bring a large community together to provide widely accepted interface methods.
You might look at this as a German and a Russian, neither of whom understands the other, communicating by mutually agreeing to read and write in Spanish.
The problem is, of course, that the information does not translate cleanly. The German has words for things that don’t exist in Russian. Putting the Spanish translator in the middle just muddies things up more.
Back in the insurance world, companies don’t have a common ground for the data they keep. Information required by one company is different from another. As a result each vendor has to conduct extensive tests with each company before their interface can be delivered. This means that each vendor has to essentially work out a custom interface with each company, even though they all claim to be following the same standard.
To make things worse, some of the companies are reluctant or unable to even share this information (or “edits”) with the vendors. Any vendor attempting upload is likely to discover that keeping up with changes in the company system is a major challenge.
In turn, the vendors don’t keep data in the same way. Some systems even allow agents to define their own screens, which means they have no standard at all.
Do you see why this is taking such a long time?
Lesson Number 5: When a vendor or company tells you that they follow the AL3 standard, that almost can’t be true beyond a certain level. Very few, if any, company/vendor interfaces are based solely on the standard as published by ACORD. That’s why it’s so hard to bring on new companies and vendors.
I certainly don’t want to hinder the development and implementation of the electronic interface efforts. They’re an extremely worthwhile pursuit, and show a lot of promise. To be sure, this kind of technology is especially valuable to large agencies with a lot of data.
But realistically, I don’t see how the average agent can justify the cost of this effort. The same might be true for many small and medium-sized companies, too. (We won’t get into the fact that almost half of all agents don’t even have an agency management system at all!)
So what’s the average agency or company to do?
I’ll tell you what one system vendor is going to do. Doris, Inc. is going to watch and see how all of this goes. Is it AL3, or is it VisionWare? What about the new AT&T efforts?
We’re going to attend the ACORD meetings and the VisionWare meetings. We’re going to talk to the leaders to learn what progress is being made, and how the average agency or company can put it to work.
But we’re going to challenge the claims, and ask for specifics all along the way. And we’re going to share everything we know publicly.
Finally, we’re going to try to help out any way we can to see if there’s a way to bring this technology to the agent on the street at a price the agent can afford. Maybe that means embracing more than one technology. When we low the answer, then we’ll move on it.
We honestly don’t know right now how this will turn out. Probably some form of AL3 will win out, because so much time and money has already been spent in developing it. But it has to get easier before all of us can share it. Companies have to come together to find a common ground. Vendors have to find a way to implement this technology in an affordable agency management system. And agents have to buy into the concept of faithful and accurate data management.
Or maybe this is just for the big guys.
Ken Butler is president of Butler & Associates, a national consulting firm specializing in the insurance business. He is the author of several policy administration systems, as well as the DORIS agency management system.
Copyright Rough Notes Co., Inc. Aug 1994
Provided by ProQuest Information and Learning Company. All rights Reserved