In the intricate world of statistics organizations, a common challenge often arises: how to balance increasing cost pressures while maintaining the quality and relevance of the produced tables. In this episode, I dive into this recurring scene, shedding light on the art of negotiation on the numbers of tables.
I also talk about the following important points:
- Understanding the Numbers: A Prelude to Negotiation I emphasize the need for innovative solutions to rising costs in table production, stressing that understanding the essence of negotiation is crucial in a world where numbers matter.
- Stories of Success: Reducing Tables Through Strategic Thinking Through two stories, I illustrate the potential of negotiation, including reducing tables and restructuring them to create savings.
- The Art of Negotiation: Key Insights Active listening is a powerful tool in negotiation that helps negotiators understand each other’s needs and concerns to reach mutually beneficial agreements.
- Leadership Skills: The Cornerstone of Success Investing in leadership skills is necessary for teams striving to reach excellence.
- A Beacon of Excellence: The Leadership Training Program Gary Sullivan and I pioneered an innovative leadership training program that provides professionals with the necessary skills for success in negotiation and leadership.
- Modular Table Production: A Cost-Effective Innovation My modular table production concept strategically reduces costs, increases efficiency, and ensures consistency in the production of repeat tables.
Do You Still Negotiate About Numbers of Tables?
[00:00:00] Alexander: Welcome to another episode of The Effective Statistician and today it’s just me, Alexander. There is a recurring scene that is happening and that I’m seeing again and again across different companies and for many many years. And the story always goes in the same way. On one hand, there is the statistics organization, and the statistics organization is working a lot on producing tables.
And these tables have increasing prices. And the statistics organization pays a lot for either external vendors or internal resources to create all these tables. And now there is increasing cost pressure. And if you think this is a recent development, no, it’s not. We have these discussions for decades to cut costs, be more effective, find new ways, all these kind of different things.
At the same time, I see prices for producing tables going up. I’ve recently seen prices per table of more than 3, 000. I can still remember when, you know, there were unit prices for repeat tables of less than 100. But yeah, that’s, that’s another topic. Yeah. And so there’s a fight between the stats organization that wants to. decrease the number of tables and the people from regulatory, from medical affairs, from HTA, from PRA, from clinical, from all these other areas that want more tables.
And this story comes again and again and again. Does it? Is it familiar to you? So I’m just thinking back of some time when I was working for company. And they had a new goal. They wanted also to reduce the number of tables. And since they said for each study, you can only have 250 tables. I think it was about this, this number.
Okay. And now there was a colleague who had a list of tables that were more in the range of 800, 900. And that colleague Now I had this new goal, this new KPI, Key Performance Indicator, of only getting 250 tables. And of course this person achieved it. Guess what she did? First, she had a discussion, of course, with the counterparts, regulatory people thematic.
And what can we get rid of? Oh, and maybe some tables were not as important. The vast amount of tables she got rid of. by just renumbering them. Yeah, you know, there’s very often these tables that you do by age and gender and pretreatment and whatever other subgroup. Yeah, or you have tables for the questionnaire for the total score and for some sub scores, things like this. And of course in her first counting, every single was an individual table. And then she said, okay, everything that is by total and then by subgroups We make that one table. Yeah, and you just had further rows in there that’s that by total and by subgroup XYZ. Or the ones that were for total score and sub score.
A subscript B Subscript C. Yeah, something that you very often see, for example, in neuroscience or in areas where you have very kind of complex diseases, think lupus or some of the other immunology diseases, things like that. Yeah. You have these. All these sub symptoms, or these special areas, and so there’s a lot of way to basically repeat more or less the same table just for a different endpoint or for a different subgroup.
Same of course by week or, you know, there’s so many kind of different things that you can think about how you can organize more or less the same content. The same analysis just with different endpoints, subgroups, and so on. Or you have, for example, different estimates. Yeah, you have one analysis that is policy, one analysis that is on treatment, another analysis that is X, Y, Z. And you have that for similar endpoints again and again. You could combine these into one. Yeah. Ultimately, you could combine probably nearly everything into one table. Yeah. Okay. So that’s what you get. She got below the 250 threshold, job done.
Another story. I was recently at a company and providing some training and this company isn’t special from all the other com companies, you know, big pharma company. And I see this topic again and again from others as well. And. During the trainings, I said, we always run into the same problem. We have these database locks and we give a certain number of tables that we want.
Yeah, like we only give 20 tables or 30 tables or 40 tables, something like this. And the people that we work with, again, common suspects outside of statistics, they want 60 or 70 or 80 tables. And so we have this discussion about how many tables. And of course, we statisticians, we get pressure from upper management, reduce costs, not so many tables.
And the other said, yeah, but we need this information. Really, we need more tables. Again and again and again. And how do you get out of it? That is really the problem. We have these established systems, that we count tables, that we pay zeros per table that we have these standard tables, all these kind of different things. And within this system, it’s really, really difficult.
One other aspect of the system is, of course, that the people that you work with, the people in regulatory or the people in medical or whatever, they usually don’t pay for these tables. Yeah, I’ve yet to come across a company where there’s any kind of internal charging. Oh, yeah, I want these. 3000 tables and yeah, here is the money that I pay for this.
Actually, I know one company where that is the case. So, you can think from your stats organization really, really careful about what do you get in return. And that is not kind of a rhetorical question. No, really think deeply about what you get in return. And we’ll get to this a little bit later in this episode.
Obviously there’s also a problem here with outsourcing because if you outsource everything and you pay per table, then you have a, you have a problem there. Yeah. Why? Obviously the first problem is that the table is not a defined entity. What do I mean by this?
Have you ever thought about what exactly is a table? I can tell you how long one meter is, and we all agreed on how long one meter is. We all know how heavy one kilogram is, or pound, or mile, whatever you kind of system you work in. We agreed on that. We agreed on how long an hour is. We agreed on… What is a euro or a dollar or all these kind of different things?
We agreed on lots of lots of different things. I’ve yet to come a definition of a definition. What exactly is a table? So the first problem is you negotiate on something that is actually not clearly defined. Now I said the first keyword negotiation. Negotiation is one of the key. Key skills that you need to have to be successful in your job, probably in life, because you negotiate all the time. You negotiate about promotion, you negotiate about salary increase, you negotiate about going to a conference, you negotiate about, about number of tables.
And so all your stats skills don’t help you. You can be the most amazing, brilliant methodological statistician. You can know everything about estimates. You can know exactly the details of how an MMRM model runs. You know exactly kind of all these kind of nuances. Within your therapeutic areas, the different scales and the different symptoms and the different treatments and all these kind of different things. If you don’t have good negotiation skills, you’ll not succeed.
One of the key frameworks within negotiation is the understanding between what is your position, what is the other position, what is your intent, and what is the other side’s intent.
Now let me explain you a little bit what I mean by position and intent. And I’ll not directly, and let’s go to the table examples. One side says, I want 20. I will only give you 20 tables. The other side said, but I need 60 tables. These are the two positions. 20 tables versus 60 tables. What they say. That is, the position is very often kind of, it’s where the controversy is about, where, you know, the problems arise, because obviously these two positions are mutually exclusive, either it’s below 20 or it’s above 60.
Nobody can win. Yeah. Of course you can find some kind of compromise and say, well, 20, 60, let’s split the difference. You get 40. Okay. Yeah, you can negotiate that way. Yeah, but then nobody really wins. Yeah, you have actually Doubled the price and the other side still lacks 30 percent of the information that I need. Not really a good way. The concept about position is That you don’t negotiate about the position. What you negotiate about is the intent?
Now, the intent is hidden. It is not clearly there. And to explain your kind of what is the, what is the intent versus the position, I want to take you out of this table example for a moment. The typical story that is told to understand what is intent versus position is the story of two girls. Two girls fight over an orange.
Both want this orange and there’s just one and they fight about it. I want this orange. The other one girl says, no, I want this orange. And they fight about it. And then the father comes and says, please stop fighting. And girl A says, yeah, but I want this orange. And the girl B says, no, I want this orange.
And then the father, Well, these are the two positions. Both girls want the orange. And of course, this is mutually exclusive. Yeah? Either one girl gets the orange, or the other girl gets the orange. And of course, you can split the difference like usual. But, the father understands the difference between position and intent.
And now… And he uses another important leadership skill, probably the most undervalued skill whatsoever. And that is listening. And listening just doesn’t mean just sitting quietly there. Listening also means asking the right questions. So he does what good negotiators do. They listen. And if you don’t believe me that good negotiators listen.
Scroll back a couple of hundred episodes to an episode with Lucene Mirabi, one of the world’s leading negotiators that I was really happy to get on my show just before she became super you know how should I say it? Super famous. FBI.
Negotiators, these professors on negotiations, she has them all on her podcast. Yeah. She is so famous on LinkedIn and has, you know, so many interesting people there and her rate of how she’s growing her following on LinkedIn is, is really amazing. And she, you know, helps these big companies to negotiate and all these kinds of different things.
When I interviewed her, I asked her, can introverts be good negotiators? And she said, absolutely, they can be brilliant negotiators, because the most important skill in negotiation is listening. Okay, let’s go back to the story of the father and the two girls. He is now listening, and he is asking the girls a very important question.
You need this orange. What do you want to do with it? What’s your goal with it? What’s your underlying goal?
Girl A says, I actually want to produce some orange juice.
Girl B says, I actually need something of the skin of the orange for bakery.
And now it’s pretty obvious. What do they do? Goal A gets the juice. Goal B gets the rest.
Both win. You created a win win situation. Both get everything they want. There’s no splitting the difference. Because both have an intent. And both get clear on the intent. I want to use it. drinking it and I want to use it because I need it for baking my muffins whatsoever.
And this is exactly what you can do within the tables example as well. Understand what is the underlying intent. For you as a statistics function, the underlying intent is probably that you want to decrease costs. You don’t want to overwork your people. You have just maybe, you know, five people that can work on these tables in this given time frame.
And you don’t want to have them work on the weekend and these kind of things. Okay, this is your intent. These are your constraints. Now let’s look at the other side. The other side doesn’t need the tables per se. What they usually need is the information within the tables to make certain decisions. When do they need to make these decisions?
For what? What exactly do they need? Do they need figures? Do they need exactly the tables? In which format do they need it? All these other things, yeah? But very often, they don’t need 60 tables. They need the information in these 60 tables, and they need it for a presentation whatsoever.
So, have a discussion about what is the intent. And then, now it’s not as easy as with the orange example, you need to find creative ways to solve this problem. And now there’s a, there’s another important leadership aspect. And that is understanding the business. How does the business really work? What can I really do?
How do I for example create an rShiny app? Or how do I negotiate with a vendor? Or all these kind of different things. What can you do? To cut costs, maintain your resources, maximize the effect of the resources you have to get the information in the right format that people can work with it at the right time.
What is the other timing? Yeah, maybe there’s also room. Yeah, maybe the other time a person is, yeah, actually we need these then because then we give it to our external vendor and they need 20 days to create these figures from it that we need. And you can say, hmm, actually we could create these figures directly.
Then we have 20 more days. How does that sound? And actually you don’t need to pay these other vendors for these figures that cost, I don’t know, 5000. K per, 5K per figure or more. Hmm. Okay. Understand the business. Listen to what is the intent. Find creative ways. Negotiate. All these things.
And when I talk about leadership skills, I don’t restrict this to supervisors. Obviously, this is, I never have a talk here about having a direct report, being a supervisor, being a manager, all these different things. No, this is the same at every level, whether you have direct reports or you don’t have direct.
You need to have all these different skills. Let me give you an example of how it could work differently with the tables. Back in the day, I was working on a HTAE submission in Germany. And maybe you have listened to this. So story here or there on one of the hundreds of episodes here on the podcast.
If you haven’t, here’s the story. There, with the German HTA system, you really need to provide lots of information and actually lots of tables because these tables are predefined format from the GBA, the political decision maker in the German healthcare system. So it is really about getting a good price for your new drug, for your new indication of your new of a drug.
And in this case, because we had so many subgroups, so many endpoints, so many different populations, different ways to analyze the data, we came up with thousands of analysis that we needed to provide. Yeah, just, just give it a number, let’s say 10, 000 tables. If you think about 10, 000 tables, and you would pay for each of these tables 1, 000 euros, that’s 10 million euros.
Of course, this is an enormous amount of money, and I didn’t want to spend this money. Even if I would have it, I wouldn’t spend it, because that is ridiculous. Because most of the tables are really repeat tables. And very often you have, and in the situation we actually had, the situation that you need to change a little bit, that affects lots of lots of different tables, and it’s actually not a lot of work. If you have created your tables in a good way, in a modular way.
And then you need to rerun all these tables. Do you need to pay then another 10 million? We actually had that, yeah? We got the feedback, no, your population that you looked into is wrong. You should exclude all these different patients and then you can make your case. Oh my god, we needed to rerun thousands of tables and we only had days for it.
The good thing was… We didn’t pay per table. We set up a very, very good working relationship with the CRO. We had lots of face to face meetings with them. We had lots of interactions with them. We exactly, you know, went through all the different programs. We explained about the background and all these different things. And so there was a very, very high trust.
And that led us to get All these thousands of tables update within three days for, I think it was, I don’t know, two euros per table on average. Two, not 2k. So we had really, really fast timelines. We had really High quality. We got a great overview of what has exactly changed. Where are the p values that are now significant or not significant anymore? Where do we have a different benefit rating according to the German system? We had all this within three days.
How? Because we didn’t negotiate about number of tables. We negotiated about what exactly do we need. And the other side, well, we, you know, the CRO said we need to contain, of course, cost. We need to make our cut. All these kind of different things. We listen to each other. We understand what was our difference in tens.
We helped. Each other understands the business. They helped us understand how the programming works, what they can offer, all these kind of different things. We told them a code, here are the different timelines. These are the things that are non-negotiables. For example, the format of these tables, because that was given by the HTA body.
Yeah, so we are very, very clear about the intent. We listen to each other. We were sharing our understanding of the business and we found creative ways to make this happen. So as a summary, in order for you to achieve excellence, to cut cost, to Increase quality to increase speed what you need. are leadership skills.
You need negotiation skills like the framework I talked about, intent and position, listening, understanding the business. You need these to be effective.
Now what do you do? Now you know, okay, you need all these listening skills. Now there’s a couple of different ways how you can increase Your leadership skills. One is, of course, you could go back to all the different episodes that I have here about leadership. I have many of them on the podcast. Second is.
You probably have a negotiation class, a presentation class, yet another whatsoever training within your organization. And you can go together with your sales people, your medical people, your people in IT and all these kind of other areas. You can go together to these trainings. And you learn a little bit here, and you learn a little bit there. And bits and pieces you, you know, put everything together over the next decade.
And of course then, you don’t get something that is specific for your needs. You don’t get something that is specifically for statisticians. You get a general leadership training. Which is not bad. However, what I have experienced, it is really most helpful If people have a similar background, and of course we statisticians, data scientists, programmers, we have a similar background.
We have typical situations. We have a common skill set. We are very often very detailed, introvert, all these kinds of different things. And in order to help you, Gary Sullivan and myself, we have put together A leadership training that has evolved over the last 12, 13, 14 years and is now in a shape that has helped hundreds of statisticians.
And we regularly get very, very good feedback on this. Yeah, so, at the end of the trainings, we ask participants how likely would you recommend your training on a score from 0 to 10. And the average is always about 9. 9 out of 10. How likely would you recommend it to your peers? So this is not my feedback, this is what other participants have told us about this training.
Now you can get this training in two different ways. One is… You can contact Gary and myself, or myself, and we can have a discussion how we deliver the training within your organization. Yeah, and we have various big companies that have been requested this to us and then say, you know, we train 30, 40, 50, whatsoever, or 20, you know, what is the number of statisticians, programmers, data scientists in this area.
Or, you’re just alone and you want to have this training for yourself. And you can go to your supervisor and say, hey, can I get this training? Then, because it’s an interactive training, yeah, it’s not just, you know, online listening to videos. No. At the moment, we offer this training again for you to join and to get through this training.
Get all these learnings about leadership skills specifically designed for you. And the training goes over several months, and there are nine group discussions where you know, maximum of 12 people come together and we talk about the different things that we have trained in the videos. So it’s a hybrid.
Or how you should say it, blended training, you know, consists of self-learning and group discussions. If you’re interested in this, you have the opportunity up to 15th of November to enroll into this training. Actually 14th of November, because it starts on the 15th of November. If you want. Check
theeffectivestatistician.com Find their courses and get to the leadership program. The leadership with masterminds. Masterminds is what we call these group trainings. And then you can just sign up. So, invest in your leadership skills. Because this This is the one thing that you need to achieve excellence.
Whenever it is about working together, you will need these leadership skills. Having SOPs is all fine. Having MSAs with your heroes is all fine. If you want to get out of these never ending, reoccurring discussions, you can only do this by gaining new skills, just trying the same thing again and again will not work.So do yourself a favor, invest in your skills, invest if you’re a supervisor, invest in the skills of your team, of your department and increase your leadership skills.
Never miss an episode of The Effective Statistician
Join hundreds of your peers and subscribe to get our latest updates by email!