nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Customer Data

Jason_Wiggins

Hey All,

The Rocky's team have had several circumstances where our customers want to share data with us to help

-Structure a JMP day that is relevant to the current and prospect user base at the company

-Help customer solve a specific problem

Could you share your strategies on how to work with this request? One of our challenges is that getting an NDA may take a lot of time and ultimately create a roadblock to helping our customer and growing our user base.

Thanks!!!

Accepted answers

Phil_Kay

Hi Jason,

I avoid using customer data.

The main reason is that our company policy is we can’t receive customer data without an NDA in place. And you are not going to go through that process with most customers.

IMO you wouldn’t want to use customer data anyway.

You don’t understand the context of the data like they do. And context is everything.

JMP is a volume business. As an SE I don’t have time to customise my presentations and demos to every customer. I use stock demos over and over again. This enables me to present to more customers. And to refine and continually improve my presentations.

Our stock demos are better for demonstrating the value of JMP.

If you are presenting to a large audience then a specific dataset from them will probably only be relevant to a subset. Better to use an example that any scientist or engineer can relate to.

I would spend some time letting them explain the kind of data they need to use, the challenges they have and the value of being able to get more from this data. Maybe they can share an anonymised example to help you understand. Then explain why you are going to use one of our stock demos.

Hope that helps.

Phil

MikeD_Anderson

Maybe a different take on this issue than what Phil has provided.

There are lots of opinions on this in the SE community and JMP in general. We went down this rabbit hole in the US Mid-Year Technical Summit last year. It was a good discussion. Some teams are comfortable with the extra steps required for touching customer data, some don't want the trouble. I will say the importance and frequency of occurrence for this issue has a lot of dependence on the territory. If your territory has a lot of business coming at you ("shut-up and take my money" situations) you probably can avoid any customer data with some careful qualification. Other territories are more prospecting driven and you have to be more open about this. So, there are a lot of moving parts in this question.

Phil is 100% correct in his first statement - we should avoid it whenever possible for legal reasons. And when we can't, we must get an NDA. This is for our protection. Also, our customers sometimes don't know their company's policy around data sharing and are more open than their lawyers would like. So you have to protect them as well as us by taking a cautious line with customer data. (Oh, and we don't accept data with personal info - period. Even with an NDA.)

You've also got to be careful about it from the standpoint of crossing the line from "selling JMP and supporting the user's growth" to "consulting and professional service provider." This is particularly an issue when you are viewed as a trusted advisor by those that you support. We can't consult. We can demonstrate the art of the possible with JMP - but we can't consult and there is a fine line there. If you find yourself shifting to consultant it's time to start pushing those requests to an internal champion or involving one of our partners.

All that said, I generally let the needs of the customer drive in these cases, and it really does depend on their needs whether interacting with a customer data set is truly required. When it is, I've been trying lately to first get them an eval copy and look over their shoulder via webex rather than accepting data - talking them through what they are trying to do. This requires a lot of patience on our part - but it helps them up the learning curve and also makes them confident in their ability to do it later. I've been finding a lot of success with this method - particularly with Pro opportunities.

You can also ask them if they have (or know of) publicly available data that is in the same vein of a problem that they are trying to solve. They may know - or they may be able to give you enough info to find something that will resonate with the audience. With enough information, you can probably find something on Kaggle or the technical reference library that will scratch the itch. This worked for me recently - a 15-minute phone call with some detailed Q&A helped me find a data set that would give them what they needed.

You can also have them use the table anonymizer in JMP. People forget about it, but it's really useful for these situations. Just have them run Tables > Anonymize Table in JMP with the data table open. Ask them to rename the column names or group them so you know what are factors and responses. At that point, they know what the columns and any categories are but to you it's just columns of numbers and anonymous categories. It makes the analysis a little trickier but sometimes that will be enough.

Now, assuming they are insistent of seeing their data used with JMP. They don't want to use the anonymizer. They want you to drive or it's in a group demo setting. You've made sure all the legal boxes are checked. You need to make sure that they understand that their data may not be able to show everything that they are asking to see (too small of a set, too messy, etc.) and that you'll bring in data that you have to supplement in those situations. If they're okay with all that and are still with you - I'd go for it. First - by this point, they've got skin in the game so odds of a sale are higher. Second, at least in my area, this would be a Pro situation at this point (I've been able to handle everything on standard JMP sales with the other methods). So, if it's a choice between a Pro sale (with customer data) or no Pro sale (without customer data)... I'm not inclined to leave money on the table.

I'm generally pretty excited if the deal gets to that point, too. The biggest deals I've been involved with (exclude large buys due to acquisitions or miscounts on user numbers) have got to the point where we needed an NDA and are involving customer data in some fashion. They are usually multiuser Pro deals where there are some serious pains in the company and they need to see if JMP Pro can help them gain ground. These situations are generally, at least in my patch, very high ROI on my time and have a high success rate. But as has been said - you've got to be really - REALLY - careful about the data - even at this point.

If I could decouple the discussion of "custom demos" from "customer data demos," I believe very strongly, and particularly when you're starting up as a SE, that you should customize every demo you do (customer data or no). You'll learn more about JMP and you'll get a better feel for your own particular demo style and how you like to organize them (journals, power points, etc.) if you're starting from scratch every time (at least at first). If you are dogmatic about "Great Demo" every demo should be customized to give the customer the biggest wow moment possible. You can do a custom demo out of a stock journal - but it's always - always - custom. I've got a set of stock demos, most SE's get to the point where they do, but I'm always adjusting or making copies and modifying them for the needs of the customer or the industry that's going to see them next.

So - to sum up:

Avoid accepting customer data whenever possible (this will get easier as time goes on and you flesh out your personal collection of "demo safe" data sets)
Try to have the customer identify a public data set or provide some detailed understanding of what their data looks like so you can find something nonconfidential.
When it's not possible to avoid, make sure you're protecting the company and your (potential) users (NDA's, anonymization, etc.).
If it's a choice between customer data and a sale - take the sale. Particularly when the ROI is high. (being sure to follow point 3, of course)
Always customize demos - especially when you're just starting out.

Best,

jiancao

These are very thoughtful comments by Phil and Mike. Here are my two cents:

I'll avoid using customer data for my first demos but find a data set that is as close to their analytical situation as possible based on the info from pre-demo calls. For subsequent presentations I'd take their data only if (1) the deal is large, (2) the objectives are well-defined, and the data is clean (relatively speaking), i.e, ready to go into JMP, and (3) a firm commitment to licensing if the stated objectives are met.

All comments

Phil_Kay

MikeD_Anderson

Maybe a different take on this issue than what Phil has provided.

So - to sum up:

Avoid accepting customer data whenever possible (this will get easier as time goes on and you flesh out your personal collection of "demo safe" data sets)
Try to have the customer identify a public data set or provide some detailed understanding of what their data looks like so you can find something nonconfidential.
When it's not possible to avoid, make sure you're protecting the company and your (potential) users (NDA's, anonymization, etc.).
If it's a choice between customer data and a sale - take the sale. Particularly when the ROI is high. (being sure to follow point 3, of course)
Always customize demos - especially when you're just starting out.

Best,

jiancao

Jason_Wiggins

Thank you Phil for your thoughtful comments!

Jason_Wiggins

Thank you Mike for your thoughtful response! This is very helpful!

Jason_Wiggins

Thank you Jian! This is very helpful!

Phil_Kay

Re: help the customer to solve a specific question...

My favourite approach (and Jon Weisz recommended this to me as well) is to ask them to post it in the community. Ask them to let you know when they do so that you can reply if you are able to.

This is good because...
It makes it clear this is not formal consultation.
It brings them further into the world of JMP.
They find out that there is world of JMP users there to help them.
Less work for you because often someone else will answer their question before you get chance, especially if you are on the road.
They often get a range of suggested solutions so find the best one for them.
Other community users can benefit from the responses.

If they don’t want to post to the community then it is likely that they are not sufficiently motivated to invest their time to find a solution (but they would happily waste your time). Or it is not something they should be sharing publicly, in which case they shouldn’t be asking you without an NDA.

malcolm_moore1

I agree with the comments in this thread. Try and avoid use of customer data at all times, it is rarely needed to advance an opportunity, and certainly do not use customer data without an NDA.

Also remember to ask why? Why is the customer requesting we show something using their data? Try and get to the real reason they think this is important. You may then be able to make an alternative proposal that is more efficient of our time and advance the opportunity faster.