Create a prediction market for your MR project in only 20 minutes
There are no companion slides for this presentation
Create a prediction market for your MR project in only 20 minutes
Presented by Hubertus Hofkirchner, Chief Futurist & CEO, Prediki Prediction Markets, Austria “How do I design a prediction market for my own Market Research project?” “How long does it take? What do I need?” Join this session and learn – live – how to create your own prediction market, tailored to your specific project, start to finish, in only 20 minutes on Prediki.
Transcript of recording with Hubertus Hofkirchner – generated automatically by HappyScribe which means it will be about 80% accurate – if you spot confusing errors, please email ray@new-mr.com. The timestamps are included to help you jump directly to a point of interest.
[00:00:06.610] – Hubertus Hofkirchner
Hi, everybody. So you’ve always wanted to use a prediction market for your market research project, but you weren’t sure how long it would take, what you need and how to actually do it. So today we’ll do one in 20 minutes. I’ll show you as a first step what you need to get going. Show you a typical concept test in the prediction market today. All you need for that is a stimulus of the intended product here. I’ve invented one.
[00:00:35.470] – Hubertus Hofkirchner
It’s called Blockbuster and gives a couple of words about what it does. It helps against nasty viruses description a couple of reasons to believe a price. And if you have it, an image of the intended product, you also need a competitor. I’ve used the bar, which is very frequent in the United States, the kind bar grain’s bar. Exactly to the same format. And you also need market data. Recent recent ones I’ve already collected here into Sum’s, and you can see that client is currently number three in the market with a certain market share.
[00:01:17.140] – Hubertus Hofkirchner
And last but not least, you will need a prediction market. Of course. Please show this with our prediction market. You don’t need to own installation or server these days. You just put it on Sarsae so you can start your own prediction market by clicking here. First thing is you have to buy enough seats for your study. You will need a plan because only public questions are free essentially for ten users. But for a market research study on a prediction market, you want at least 30, 40, 50 users, 100 if it’s qualitative.
[00:01:54.220] – Hubertus Hofkirchner
So I say by now, system asks me how many users I want. 50 is normally fine. So that would cost me 250 dollars a month. I will pay just monthly because for a concept just a couple of days is usually enough, I say next. I into my billing information. And then already get my invoice, which I need to pay by either PayPal or credit card, but that’s essentially it. When I design a prediction market, I typically like to scribble it first in a spreadsheet so I can see all the questions in context, the rules which I want to apply, the answer options which I want to offer to participants and a little bit the text which I would like to see in the wiki.
[00:02:46.100] – Hubertus Hofkirchner
So concept this typically is comprised of three sections. It has a test pattern, which is called Alua for a concept test which is short for liking uniqueness, relevance and estimates. You can compare that to something like action standards in traditional market research. There’s a benchmarking session, a section where people have to first think about the market as a whole and then about the identified competitor from whom we want to steal volume. And last but not least, the hypothetical questions about our new product here.
[00:03:20.630] – Hubertus Hofkirchner
I’ve done two because we want to get a little bit of a sense of price elasticity at the same time. So I’m offering to the audience disbar at one point sixty nine and one point eighty five as a low price. And the high price deal for that type of question has to specify that it will be judged at the actual market. For instance, one year after launch, if the price is around that area, plus minus give and take a little bit.
[00:03:54.040] – Hubertus Hofkirchner
And start entering the first question, the system first wants to know what our project is called and I’m going to call this blockbuster in in the United States for 2021. And I make this a new topic and I make sure I make this a confidential topic so that our competition doesn’t know the fantastic and, you know, Leibovitch we’re planning and off we go. And we can enter the first question in the next step by simply copy paste the data from our scribble here into the pretty key new questions kept module.
[00:04:37.630] – Hubertus Hofkirchner
So I’ve already done that here. Which of these parts will Americans like more? I have Ed Hulewicz I’ve designed here. I picked the right question. Time for that. Now, normally I just actually don’t really like to do additional prediction market, which first generation systems did, which had a binary answer. Yes, no. Somebody Suyin or somebody else going to win. I almost always use alternative. So the answer scale here. I’ve designed an answer scale.
[00:05:06.340] – Hubertus Hofkirchner
That kind pass will be like a lot, a bit of both equally and the same for Blockbuster and for reference data for loud five days for this concept test to run so people can be recruited, can reconsider their answers a little bit and we already expect the results. Also on that Sunday in a week, the time zone for their sister, since this is for the United States as America, New York. Now, I can preview the question. Everything looks fine and I publish it.
[00:05:41.760] – Hubertus Hofkirchner
Now we see our question battery and how it’s populated already with the first question, and now I say add another question to this topic and I have to do essentially the same with the other questions in our battery. I will show you one more question here, the market question, because that’s a different type. So this one, for instance, will the American granola bar market or shrink in 2021? Obviously, it’s going to be judged by the actual year on year change of the U.S. canola.
[00:06:17.460] – Hubertus Hofkirchner
Bull market in real terms here, you can see how favorable this to to use the on to scale, I’ve put here the stagnation scenario into the middle and defined high. And this was a shrinkage and growth at the city maximum, which we can imagine given from recent market behavior. Obviously, this cannot be done quickly. We will need to wait for year end here until we will know the actuals, maybe even give a couple of days more for the results expectation.
[00:06:54.360] – Hubertus Hofkirchner
So we get that in January, maybe 15 days after. Now, we should make this a little bit prettier. So to keep our participants engaged for there to add a little bit of imagery for this, I going to the. Ed managed to picture and enter a couple of pictures, like some of our media where people eat snack bars and confirmed that I have to rights to this, and then I upload the image and now I can set this to use as the image picture.
[00:07:32.080] – Hubertus Hofkirchner
And so when I go out here. And, Jack, the topic battery. I can see this already being pretty and nice. Now that we’ve entered a few pictures, let’s explain to the participants what they actually have to do and show them the stimuli. So I’m going back into the edit models and we’ll just use these new pictures, which we’ve done. This is a little editor with which we can edit the data. I’ll just paste that. So so we explaining to the people what they have to do, most people who come here may not have seen a prediction market before, so we give them a tip that they need to take a third person view and predict how other people will like this product irrespective of their personal preference.
[00:08:28.810] – Hubertus Hofkirchner
And so here down here, we don’t have the stimuli of the bars. And when we’ve done that, we preview this. OK, looks good. We enter some reason for the version control system. And safeties and we are done. So our. Battery is starting to look good, and what we’re going to do next is we will invite our boss, who is Albertus at predicting to review this prediction market questions and approve them. So we recruit him for that.
[00:09:21.210] – Hubertus Hofkirchner
We will. Send him here as a moderator. And. I into my own email, HGH at pretty key dotcom, and I don’t need to add a message since I know why I’m why I’m getting this and to send myself an invitation. Once they’ve done that, the system sends an email automatically to me personally, so I cycle to my emails and here we are. I’ve already received an invitation, an invitation from predicted to join this new battery.
[00:10:05.550] – Hubertus Hofkirchner
It will make me a moderator if I choose to participate. So here’s a little bit of information, how long it will take and how much there is to win, which is not yet something because we haven’t entered incentives. So a log in here with Google, with my personal account and before the system lets me not have to confirm that I will keep this confidential. This is, of course, also important for participants. So, yes, I agree.
[00:10:33.120] – Hubertus Hofkirchner
And now the system makes me a moderator of this prediction market. Better for all the questions. It’s very important for a prediction market to have a moderator, to help participants to to guide them a little bit, to motivate them to to do more, to do everything, to trade more. Now, my first job as a moderator of this battery is whether my analyst has entered all these questions correctly. I mean, and we know he has it was me.
[00:11:03.810] – Hubertus Hofkirchner
But just to show you so I go into one question. It tells me that this question is not yet tradable needs support. So take a quick peek, whatever things are right, all the information given for participants and I like it. So I go to the support section, which is not open to normal participants. And I support this question. And from this moment on, what it’s tradable. Once I have checked and supported all questions, I’m now at pretty much my last job here, which is to make sure that participants get paid on pretty good is called sponsoring.
[00:11:40.470] – Hubertus Hofkirchner
So we go into the first question going to the whole section, which shows how the prize money will be settled. And now we tell the system how much it should be sponsored. We simply enter here the time zone of New York. We want to start now, maybe in an hour or so. Twenty hundred we want and this on Sunday at midnight. That’s fine. The number of settlement methods available. That’s actually one of the important things to watch out.
[00:12:21.270] – Hubertus Hofkirchner
That’s the capabilities of your prediction market. So this one is going to be some price money distributed. Parata according to accurate predictions and will award this with final results. And for this easy question, we’ll give them some thirty thirty dollars, which incurs a fee of nine point for us. I say next and now I have the possibility to to use this also like a campaign if that but it doesn’t really apply for market research project. I just say here, this is sponsored by Patika and that bid next page, I intend my payment details.
[00:13:01.860] – Hubertus Hofkirchner
They are stored already and here is my invoice. I can pay my credit card, PayPal, and once I do that, the incentive is shown on the prediction market page for the participants. We go back to our battery and we can see here now that the price money is shown to participants. It’s still yellow waiting to be to start in about an hour. So that’s when we will invite recruits. So last thing for me to do is to give them a nice welcome message and say welcome to our.
[00:13:41.460] – Hubertus Hofkirchner
Bug buster prediction market. And offer them help or whatever, you can send this immediately to all participants, the unknown at this moment. So I just say it now and I pin this on the top so that everybody who enters sees whatever message I give them up here. So that’s pretty much it. That sums it up. There was a how to set up your prediction market in just under 20 minutes.
