How to analyse and prove to business stakeholders that a website redesign (b2b hardware) is better than a previous one?

You'll find that looking at top line revenue alone won't provide much value to you if you are asking the question "Is my advertising spend worth it?" Typically, in advertising I want to know the Lifetime Value of the customer or the Average Revenue Per User (ARPU). Then, it boils down to : If I spend $100 to acquire this new customer and he spends $150 over the lifetime of his paid relationship with me then, yes--- my ad spend is worth it. On the other hand: I could spend 10K this month in ads. I could sign up 20K in users/customers/top line revenue. But my data will tell me I spent MORE per customer/user than they are actually worth over the average lifetime. So that's what usually answers the "Is it worth it" question in any advertising scenario. 1. How much to acquire the user? 2. How much does the user spend over the avg lifetime? If it doesn't make sense, it's not worth it, regardless of what your top line revenue tells you. You might find this blog post on cohort analysts valuable as it goes into more detail. (I have no affiliation.) https://apsalar.com/blog/2012/01/how-to-use-cohort-analysis-to-improve-revenue/

Steve Blank has already covered this in depth in his book 4 Steps to the Epiphany, as well as a video here: http://startupweekend.wistia.com/medias/tao3s8hf7l My approach is based on his outline, with a twist. 1) I try and conduct the interviews at the customers office or place of work, at minimum on their devide (computer, mobile phone) so I can ask to see how they work. 2) I never try and sell them on a the solution, but work hard to truly understand if/where the problem is. 3) Always provide guidance to the conversation, but ask open ended questions. 4) Ask questions like "What do you do 3 minutes before, and after, you do that action (or use our product)? Other tips would be - Write down the words they use. Metaphors and taxonomies are VERY important to ensuring your product is approachable. - Use 3 simple slides: Problem, Current Solutions, Proposed Solution Hope that helps.

Are you performing a large-scale customer study? Or is this just a smaller focus group gig? The primary rationale for hiring a user research recruiter would be to obtain a valid, representative sample for statistical inference. If the number of participants is small -- too small to be statistically significant -- then you cross this concern off your list. Presumably, you know who your ideal users will be. After all, they must be your future customers, right? So if you can't track them down now, then future marketing will be pretty hopless. You can probably contact them on your own without an intermediary. Another reason for hiring a recruiter might be to inject a buffer layer for impartial selection. Otherwise, you and your colleagues may inadvertently bias the selection process toward people who know you or are already too favorable. My experience in this area is not vast -- although it sometimes comes into play for larger corporate naming initiatives. So I may be missing other arguments in favor of hiring research recruiters. I'd be glad to hear those from other people so that I can learn something too.

It depends on what you mean by "Visualized Data." Do decision makers use graphs, charts, dashboards, etc? Yes of course. Are there some visualizations that wouldn't be of any value to a decision maker whatsoever? Also yes of course. You can't show me a word cloud if I work as a portfolio manager for a bank and expect me to make any decisions from it. That being said, the social media team might be able to. It all depends on the audience, and whether or not you're making patterns they care about easier to detect. If you are talking about graphs/charts/dashboards, what I would do if I was in your shoes is I would try to make one dashboard with (say) 3-4 graphs inside of it, and then I'd show them that dashboard and say something like, "You know, I've been thinking about that conversation we had. And I definitely appreciate the point you made about fluff visualizations. But I also really think there's at least a few charts that would be really valuable to you. I whipped a few together (on my time, not yours!). Do you have a few minutes that you could spare to come have a look at them with me?"

First off, I have several people I could introduce. I'd also like to know the industry you're operating in, what the data looks like, where it comes from, and how much it needs to be cleaned up if putting into a relational database, or if the better solution would be a distributed file system like MongoDB where you don't necessarily need to normalize the data. Also, if you're a startup, or if the company is well established with many existing customers and if this is for a new initiative. Assuming you're working with a relational database, which it sounds like you are, you will want to implement something like Tableau or build out a custom dashboard using Google Charts, HighCharts, D3Js, or any number of other potential dashboarding/visualization solutions, which usually involves some programming/scripting in JavaScript. There are paid solutions like Tableau (which is amazingly powerful), and then there are free/open source options. I'd be happy to talk about possible ways to architect the solution, and discover who you would need once I understand the variables more. If you're building a web application, then you will likely need someone who is also a full stack developer, meaning they can handle building the back end and the front end in addition to the data requirements. Many early startups choose Ruby on Rails (because there is a ton of open source code out there for it) with Twitter Bootstrap (modified) and in order to visualize the data, they will need to work with JavaScript. It makes sense to have this person act as initial product and to derive the insights out of the data. They're pretty much the only person who can do this anyways, because they're the ones on the data. If you're in an early stage startup, I would recommend the strongest business owner (usually the early stage startup's CEO) be directly involved with this person on communicating what value your solution brings to clients, and what they pay you for, and in brainstorming on potential features and reports. Once the solution becomes established, and many customers start using it directly, there should be a different product person interfacing to those customers over time, gathering feature requests from customers and bringing it back to the Data Scientist/Analyst who spends their time working on the data. Depending on whether the solutions are SQL or NoSQL or hybrid, there are different types of Data Science professionals you should consider: 1. Data Scientist 2. Data Engineer 3. Data Modeler/Analyst 1. The Data Scientist handles experimenting with the data, and is able to prove statistically significant events and statistically valid arguments. Normally, this person would have modeling skills with Matlab, R, or perhaps SAS, and they should also have some programming/scripting skills with C++ or Python. It really depends on your whole environment and the flow of data. In my experience, Data Scientists that exclusively use SAS are sometimes extremely skilled PhD level statisticians and focused exclusively on the accuracy of the models (which is okay), but often not sufficiently skilled to fit within an early startup's big data environment in today's world and handle all of the responsibilities you'd like them to handle described in your question. I'm am not bad mouthing SAS people as they are often the MOST talented mathematicians and I have a great deal of respect for their minds, but if they do not have the programming skills, they become isolated within a group without a Data Engineer helping them along. Often a SAS user trying to fit into this environment will force you to use a stack of technologies that a skilled Data Architect would not recommend using. It takes programming in some object oriented language to fit into today's big data environments, and the better Data Scientists are using hybrid functional and OOP programming languages like Scala. Extremely hard to find Data Scientist can also work with graph databases like Neo4j, Titan, or Apache Giraph. 2. The Data Engineer, if you're dealing with a firehose of data like Twitter and capturing it into a NoSQL architecture, this is the person who would prepare the data for the Data Scientist to analyze. They often are capable of using machine learning libraries like WEKA to transform data, or techniques like MapReduce on Hadoop. 3. The Data Modeler/Analyst is someone who can use a tool like SAS, SPSS, Matlab, or even R, probably a very strong advanced Excel user, but likely won't be a strong programmer, although perhaps they will have a computer science degree and have some academic programming experience. The most important thing to watch out for is someone who is too academic, and has not been proven to deliver a solution in the real world. This will really screw you up if you're a startup, and could be the reason you fail. Often, the startup will run out of money due to the time it takes to deliver a complete solution or in the startup's case, a minimally viable product. Ask for examples of their work, and specifically dig into what it is that they did for that solution. I've tried to cover a pretty broad range of possibilities here, but it's best to talk in specifics. I'd be happy to discuss this with you in detail. To answer your question, is it perfectly reasonable for someone to handle all of the responsibilities described in your question, if you find the right type of person with the appropriate skills, and a history of success.