Organization Design Research

Speaker A Associates. We do primarily consulting work, but we also do research. The firm started in 1977. Since about 1990 we've been doing exclusively organizational work. We've also done research. So I have a personal interest both in doing the consulting but also doing research to find out if it makes a difference or not. So while our real business is the consulting, we do the research. Now, the downside of having the research as a sideline is that so far we've only published one study. We've got all kinds of research, but most of it is still in house and white papers and so on and so forth. So our intention is to publish more, but to this point we haven't done as much of that. So what I'd like to do is run through some slides with some of the research we've got, talk about some of the research. There's one study in particular that is published that's available and I think of interest and hopefully over the next period of time we'll be able to publish more of this. So in terms of the first piece, one of the things we do when we go into an organization and we do an assessment is that we keep track of the data. So what do we find when we go in? So we get different measures when we go into an organization, one of the things we do is collect timestamps. So what are the time spans? By collecting the time spans, you can look at what percentage of the manager direct report relationships are requisite or have gaps or have compression. So we do that as kind of a front end. And you can see that what we found, we've got data on over 26,000 manager direct report relationships from over 50 organizations. And what we find going in is that requisite alignment manager exactly one stratum above in terms of the work that's done. And when we go in, we only look at the work that's done. We don't assess any individuals, we'll often do talent pool later, but the initial work is simply in terms of the work that's done. We find that the requisite situation occurs about 50% of the time, that compression occurs about 38% of the time, and that gaps occur about 12% of the time. So we found that fairly consistently over the years, we found gas have increased slightly but not a lot in terms of the research that we've done. That's kind of some baseline ongoing data about what one finds. What we find is that when we work with an organization around improving organization design, those requisite scores go from about 50 up to about 80 90% plus. So an organization can significantly improve those scores through an organization design process. But the beginning point for most organizations is not very good. One of the broad comments I have is that what we found, and we found this consistently, is that organization design in general, and I'll talk about different ways we define it. Organization design in general is related to outcome measures in terms of financial performance, customer satisfaction, and employee satisfaction. What's surprising for us is that we found in our research that one variable by itself relates to those outcome measures, and that one variable by itself is the Manager Direct Report alignment. One variable by itself relates directly to those outcome measures. And that's a finding that we've replicated a number of times in a number of studies. This is the research that's published. So the study is called Organization Design and Performance in the Global Pension Fund Industry. Keith, Annevicher and myself. And Tom. Shebelhutt did this published this in 1998. And it's in the Financial Analyst Journal november, December. And basically this was a study of the global pension fund industry. Eight major pension funds from around the world sponsored this. So in the States, there was IBM, At T. In Canada, there was Teachers and OMRS. In Europe. There was ADP. They sponsored the research. And we did it two ways. One is we developed an instrument and we had CEOs rank their organization in terms of how good the design was. We had a response rate, because Keith works in the pension field, we had a response rate of around, I forget exactly what it was, but around 70% from CEOs, which is quite amazing. What we found with that is, in terms of CEO responses, that there was a statistically significant relationship between how the CEOs saw their design and the performance. And the performance was, and this was interesting, again, Keith got incredible data. We found that there was a relationship between that and four year pension fund financial performance, adjusted for cost and risk. It's astounding, when you think about it, you think about the money pension funds have, and to get a statistically significant relationship between this general design measure and that precise, extraordinarily precise outcome measure. Ron, the independent variable there is, I'm rating my organization on a scale of one to five or something. Basically, I forget the number of questions, around 60, 70 questions. It was a liquid scale, so a six point scale was the average weight average. We then did some subdivisions in terms of sub factors that related to it and also individual questions that related to it. What we then did is we went into the pension funds that were the sponsors. So the sponsor pension funds, as part of the deal, they got something more. The something more was we actually went in and we did time span analysis and we looked at what percentage was requisite, and there were eight funds, but we only had data points for six. Two of the funds we couldn't go into, so we've got data points for six. Now, for those of you who know statistics, normally you wouldn't bother doing a statistical analysis with An Nana six. In this particular case, the correlation was about 0.9. It was statistically significant with a sample of six, which is quite amazing, but consistent with this other piece. And we're replicating keith and I are not replicating this piece at this point, but we are this year replicating the first piece in terms of the general organization. Correct. Both work. Yeah. The first one, I forget the correlation. The second correlation was about 0.9. Mind boggling. The first one was around 0.5. But both were statistically significant. This is a study and this is public information. This goes back to Rob Pierce's early days. This is VIMO. Bank of Montreal electronic banking services. And we did a study here both on employee satisfaction and financial performance. And you can see here that organization design significantly improved. So we did a pre and post measure in terms of organization design and we defined it as percentage requisite positions. That was their definition. And again, we found that this is really just a very powerful measure here. We had, in terms of the research piece, we had baseline data. So we had baseline data in terms of performance on a larger entity. And then we had specific information related to employee satisfaction of that group. And before it was kind of back and forth and it went up significantly higher. What's interesting with both the satisfaction and performance on a financial basis is that while we only tracked it three years, in fact, this has been a continuing process. So we don't have a specific additional date on that. But in terms of the information that I have around performance, that's continued. And in fact, this organization still uses this. Similarly with financial performance, same thing in terms of the design getting better. And again, we've got a baseline that we used. The organization was significantly below the baseline and financial performance and subsequently moved up more significantly than the outside world on that measure. In this particular case, we did build in some baseline measures. So it partially takes away the issue of is it just a fortuitous thing in this one organization? And the baseline is not total organization, it's total organization minus the unit. That was correct. So here it would be the total organization minus the unit. This is a study that we've done twice. We've looked at the top 2000 companies in Canada and we've administered a survey that's gone out to the CEO and the vice president of human resources. Basically, they've completed information around their organization, their perceptions of a number of different variables. And we've got several factors that have come out of those questionnaires. So they completed them independently. They looked at some similar variables and some different variables. So we looked at from the HR person, there was an assessment of human resources. From the HR person there was an assessment of correct alignment of positions. So a significant correlation between these that you would expect if you look at measures here, intermediate measures. And this is a model that we built based on the data. This is what we would predict the data would show. And then either the data indicates that or not. We're making assumptions about the directionality here, so we're using correlations and making assumptions about the directionality manager employee relationship. Employee satisfaction is measured by the CEO. Customer satisfaction is measured by the CEO. So these two are measured by the CEO. This input stuff is measured by HR, and the financial performance is also measured by the CEO. And what we found is that in terms of financial performance, there's a direct relationship between organization design and alignment of roles as assessed by the HR person and the financial performance of the organization as determined by the CEO. What we did is one question around the CEO assessment of financial performance. One question is, is this real or is this just what the person thought? So we actually got data here, public data here. In terms of actual financial performance, there was a very high correlation between what the CEO said it was and what it actually was in terms of the external data that we had. So this is a model that hypothesizes some directionality. The most important thing from my perspective is all this stuff is statistically significant. The numbers aren't real high, but it is statistically significant. And it's a pretty consistent pattern that makes its way down to financial performance. So some of the intermediate variables, you predict more, but getting the financial performance one as well. So we've done this one twice, and each time had very similar results on an independent basis, taking a look at the information processing capability of an employee and the complexity of work required in a position. So we did in this particular study, we had two independent measures. And in terms of you can see in terms of this graph that there's a statistically significant relationship between what the information processing capability of the individual employee is and the complexity of work required in a position. This organization, by the way, was requisite, had been through a couple of palette pool processes as well as getting aligned. So whether or not you'd find that high a correlation in an organization that wasn't requisite, you probably would not. I've seen a result from a colleague in Sweden to show about the same. Okay, yeah. Without it being requisite. Without it being requisite. Okay, that's interesting. Okay, so those are the specific slides. Let me just make a couple of other comments around some other related research pieces that are here. One of the things we're currently working on when we go into an organization, we do time span analysis. So we look at percentage of requisite organizations. We've also developed a questionnaire on organization design satisfaction. So what are organization design variables? It correlates with an employee satisfaction survey. So it's a little more precise, it's a little more focused than an employee satisfaction survey is. And we're just doing some research now. We've just got a chart. And we've got data on 20 different organizations looking at the relationship between how requisite the manager direct report alignment is and what the satisfaction of the employees is. And we got 20 dots on a page like those dots. And again, it's a statistically significant relationship. So we're just working on that now. We've got two studies related to information processing capability. Both of them we did with Elliot. One was a study in the federal government, and we've got some statistically significant relationships in terms of level of capability and level of role. We've also got, equally importantly, there's no statistically significant relationship. When you look at variables like gender and someone being a minority and so on, it's good. In terms of the things we didn't get, we didn't have access to as much information as we would like in that particular study, so we still don't have as much data. The second one we did, also with Elliot is in an organization where we had two talent pool assessments. We offered employees an opportunity to go through an assessment of current potential capability, and we got about 50 volunteers. We had talent pool data from two times. The deal was the employees would get confidential data, the organization would get group level data, and we'd get an opportunity to do research. So Elliot and I did that research and again, we found statistically significant relationships between the assessments Elliot and I did of information processing capability, the talent pool, the level of work, the Interread of reliability. I must admit, the results we got were not as high as the ones Elliot gets. Elliot tends to get results. Nine, these were more middling. They were statistically significant, but they were more middling as opposed to the 0.9. So we've got a couple of pieces on that as well. What we're trying to do as much as possible is find opportunities through the consultancy to do research. And we think that that's a benefit to our consultancy practice, but we also think it's good for the field. A downside is that because our main business is consulting, we really haven't put the effort into the publication yet. And quite frankly, we're not a skill pad. That's a whole skill set in itself in terms of getting published. The other comment that I would make, having seen some projects, and I guess it's partly related to my question to you, Ken, around qualitative stuff, is I think it's difficult for an academic who doesn't understand the stuff to do research. I think doing time span is a difficult thing. I think doing information processing capability assessment is a difficult thing. I think that's very difficult for an academic to just go out and do so minimally. Consultants need to be involved in that process. I think to the extent that there's better liaison, the academics can bring stuff in in terms of the purity of the research and being able to write it up in the way a consultant would struggle to do. But I think some of those consulting skills in terms of assessing those sorts of so that it's real stuff, is important. And I think that's difficult for an academic to do independently about it. Thank you very much.