Testing Hypotheses

I am testing a couple of hypotheses. This is part of the research I am doing to build a product. I can use some help.

Hypothesis #1

You need to research a topic once in a while, for your job. The topic can be broad, for example –  ‘mobile trends’ or a little more specific for example – ‘Tablet Applications in Healthcare’.

Hypothesis #2

If you want to do the research mentioned above, there are a few options open to you:

  1. Do a lot of Google searches and copy and paste the results into your favorite document editor and do more searches on the links you found.
  2. Try asking your friends knowledgeable in the space
  3. Try asking in  professional groups (relevant to the topic)
  4. Do some searching on Twitter or other social networks
  5. Set up a Google alert on your topic and associated keywords
  6. Initiate a paid research (if you can afford it) or try to find some research reports
  7. Other…
Do you have a research need? Would you be willing to discuss it with us? (we are not trying to sell anything but trying to understand the needs better).  In your past research what did you use? 

If you can leave a comment here with your observations I would appreciate it very much.

18 thoughts on “Testing Hypotheses

  1. Hi Dorai,
    I have done some research work in the past and instead of reading through articles or journals on the internet as my first option, i followed these steps. Taking the tablet application in helthcare as an example, let me try and spell it out.

    1. Make assumptions about the users in this domain. “the users need a tablet because of the following reasons…” these assumptions could be bizarre.
    2.Now, test these assumptions. Get to the field,talk to people, in this case the hospital or a pharmacy or any space related to health care and see how the users are. How do your assumptions hold in front of how the reality really is. Assumptions have given me a starting point and have thus proved helpful.
    3.Having gathered knowledge about the behavior of the users, their enviornments and their needs, i would now go forward and read up all that I can to get a better understanding of things in that specific domain.

    Doing so has given me a lot of insights into how the product should be finally shaped.

    1. Thanks a lot. That really helps. What I am trying to validate is the need for research in companies. What role were you playing when you did this research?

      1. I was the researcher ( a mobile phone company, take a guess. Yes that one! 🙂 ) I was responsible for running into villages in Karnataka, speak to the farmers there and understand how their problems could be solved by this company.

        Research is important no doubt. In my opinion it needs to be done after a prototype is made (the product of our assumptions) and can be tested with the users. In this format research is closely knit with user testing. Doing this helps validate our assumptions better and promises better quality of product before it goes into production.

        1. There are two schools of thought. One thinks that you should simply talk to people, understand their problems and design a solution. Another one seems to think of parallel development of some minimal version along with conversations. I agree with the power of prototype/mockup. It makes the idea a bit more concrete. Otherwise it may remain abstract to the audience.

  2. While working at a startup, much of the research we did for industry analyses and management presentations was pulled for free from around the Web. It would have been super useful to have a “high quality source” filter available that removed the junk/noise or purely opinion content and rather focused on those sources that had some real data and insights behind it. I also found it relatively easy to get access to the larger, more in-depth research put out by Investment Banks, which were ripe with trends, projections, market sizes, etc.

    1. Thanks. I agree about ranking the sources. We currently use some simple ranking based on Google ranking of results technorati rank and planning to use post rank. We have a few other techniques we want to test and talk about.

      I agree that ranking sources is an important step. Have some ideas in that area. Will include it in a future post.

  3. When I was with Sony Research labs, I was asked to research and produce a report on the state of the art of ‘Molecular Electronics’ as an aid to their decision to invest in the emerging nanotech in 2001. I had little idea on this subject when I started. I delivered a good report in a month with almost all the research coming from your item 1 under hypothesis 2. Sony purchased a paid report on the subject a couple of years later and I found it to be so poorly done to be of little value. Of course, in 2001, your items 3 and 4 didn’t exist yet. So, I would think researching thru Google in my experience has been a great tool even if we lack practical experience on a subject.

    1. Thanks Raghu. I remember that research and our discussions. So you think there is a need for a tool to help research?

      1. Yes, I think a lot of us use research using the web to learn as well as augment some other work that we do. So, a specific tool that integrates the 6 points you have enumerated would be great. If possible, it should update the results with new research data as the days / months pass so long as the ‘research project’ is set to be ‘live’ by the user.

        1. Thanks Raghu. Exactly. See Sean’s comment. I am in full agreement. The real value of the tool (besides aggregating and filtering data) is to suggest new keywords to track and finding new entieties. .

  4. I think the need for a tool starts with the creation of the first set of results. How do you automate the search process so that as new companies/people/solutions/technologies get mentioned you add them to your landscape, or at least add them as candidates for additional analysis. I would take more of a wiki approach where each version / iteration of the results is available and can be diff’ed so that you can see the evolution of the analysis (and the size of the relevant landscape).

    We do a lot of research. I think the opportunity for a new tool starts at step 5; what else can I set up besides a Google alert to operationalize my insights into ongoing monitoring. “What’s new” given an existing taxonomy is one place I would suggest that you start.

    1. Sean,
      Thanks. You are right. Any tool has to be better than Google alert and need to continuously update the data but also come up with suggestions for new keywords and companies/products/events to monitor. One thing is becoming clear. As an intermediate MVP I wanted to validate the need. It seems to be there based on some of the feedback I am getting here and a few conversations we had when I was in US.

  5. Dorai,
    without having a practical idea on the topic i research ( which means the possession of just the theoritical knowledge) how can i go further and say that my analysis is perfect(approx)? thats were i get get stuck always. I am a student and i lack practical knowledge on many technologies where as “the google” has made me strong in the subject theoritically. If i want to research on a product then am i fit enough to kick start with just the theoritical info? help me 😐

    1. It depends on the purpose of your research. For example, if you are researching ‘Cloud Computing’ to understand job opportunities and whether there is a fit, the research is pretty simple. You find a job site (like monster.com) or job aggregation site (like indeed.com), search for jobs on Cloud computing. You look at job trends to see whether the jobs are increasing, decreasing or flat. You look at how long the jobs are listed (listing date) to see whether they are filled (demand vs supply). You confirm what you find from a couple of other sources like doing searches on ‘job trends for cloud computing’ make sure that your results are recent (in the past 3 months or year) etc.

      Most of the research professionals research topics that they may not be familiar with but with a few searches and some amount of reading can get reasonably familiar in a few hours or few days about the lay of the land.

      For students, the story is different. You may be looking to do a project, looking for a company to do a project in, trying to understand the space to write a paper etc. Once you know why you are doing research and depending on how familiar you are with the topic, the techniques you use and the searches you perform will be different. For example if I do not know anything about Semantic web and try to understand what it is, I may start with ‘semantic web 101’ or ‘a tutorial introduction to semantic web’ etc.

      Gathering data is one thing. Analysis is a different matter. For analysis you may need more skills. If you don’t know what skills you need, the easiest is to find a course and look up what they teach in that space or look at jobs and see what they ask for.

  6. Hi Dorai, I have done this kind of research for healthcare industry. I have followed the same tsteps but the order is different.
    1. simple google search for my topic
    2. from the top 10-15 posts I have identified my sub sub topics and repeated step 1 and analyzed the trends, alerts, news, reports every thing related to them.
    3. From my observations through above step I have verified my understanding based upon my assumptions/speculations with reference to some standard report or with my friends.

    Most of the times I relied more on Google search, trends, news and reports. I never used FB or Twitter trends, I need to know a lot about the SNS before I use them for research.

Comments are closed.