Help November 6, 2020

Ask IH: Doing things which don't scale and it's taking forever. Am I doing it right?

tsk

Hello IH,

Some background about what I've been working on. Appsolite is an app research tool. The primary purpose of the tool to is to surface what a software application is good at by bench-marking the features & functionality across the whole category(s) that app in question belongs to . I've done a few user interviews (also my personal experience working for large corps) and it seems like there is a good product-market fit during software evaluation & procurement stage.

To that end, I feel like I am trying to boil the ocean and it has been overwhelming with the amount of data I need to sift through. I decided to focus on single category (MVP) and build a report but even to do that it is taking me forever.

My question is, am I doing it right? I think this is an occupational hazard being an engineer that I can code my way out of this schlep work.

  1. 2

    What problem does Appsolite solve?

    Even if what you are building isn't intended to scale, it should solve someone's problem, at-least yours.

    1. 1

      I will just copy-paste from the other comment:

      I have interviewed a few users and it seem like software evaluation is a very resource intensive step even with industry analyst report and given the risk involved people usually go with top product in a category even though it may or may not meet all their requirements.
      There are 2 ways businesses evaluate software applications. First is using industry analyst reports (Gartner, Forrester etc.) and the second is "yelpification" where there are a lot of review websites which have propped over last decade or so. Both these approach favor a "fat head" totally ignoring the "long tail". With Appsolite I am focusing on the long tail of a product category.

      1. 1

        Okay, if there is a problem and if your product solves them effectively you would be able to sell it to those who face it irrespective of all other variables.

        As far as Doing things that don't scale, it helps tangentially to a business but never directly within a particular business.

  2. 2

    Like @csallen, I'm also a bit lost. What exactly are you building?

    Maybe you can clarify the following paragraph?

    Appsolite is an app research tool. The primary purpose of the tool to is to surface what a software application is good at by bench-marking the features & functionality across the whole category(s) that app in question belongs to .

    It could be that this is also the trouble your potential customers are having. Can you simplify that description? Maybe boil it down to just a handful of easy-to-understand words?

    1. 1

      I commented on @csallen 's question in detail which hopefully explains the solution I am working on.

  3. 2

    I'm a little lost. Who are your customers? What are you hoping to sell them? What have you built yet versus not built? What data are you sifting through, and why is it taking forever?

    1. 1

      Target Audience:
      At this stage, I am focusing on people who do software evaluation and procurement for their department/organization.

      Pain point:
      I have interviewed a few users and it seem like software evaluation is a very resource intensive step even with industry analyst report and given the risk involved people usually go with top product in a category even though it may or may not meet all their requirements.
      There are 2 ways businesses evaluate software applications. First is using industry analyst reports (Gartner, Forrester etc.) and the second is "yelpification" where there are a lot of review websites which have propped over last decade or so. Both these approach favor a "fat head" totally ignoring the "long tail". With Appsolite I am focusing on the long tail of a product category.

      Product:
      At this time, I am trying to sell reports & insights about each individual software category.
      So far I have:

      1. Database of products (100k) and categories (800)
      2. Feature set of products in 5 categories
      3. Analytics for 2 category (Help Desk Software & Electronic Data Capture)

      The thing which is taking a long time for me is to capture all features for a product then normalize features across the whole category and then creating a ranking system for each feature across the whole category. I have done this for couple of categories and the results are quite promising.

      There are 2 ways I can speed up, one is to hire people on mechanical turk or feed the data into a deep neural network using one of the existing language models like BERT or GPT2 but to do either of these things I need clean dataset to start with. Probably I overestimated my capabilities and at the same time underestimated the effort. It's taking me a long time to create these clean datasets.

      1. 2

        Are you sure that quantity will solve your problem?

        Here is what I would do:

        1. Find a company which is currently evaluating which software to use for one specific need.
        2. Convince them to pay you to evaluate the software for them.
        3. Turn the data you gathered while evaluating for them (and getting paid!) into a product.
        4. Rinse and repeat.

        Advantages:

        1. You get paid while and for building your product.
        2. You get early feedback and understand what your customers are looking for.
        3. You are spending your time on things which people really need.

        Future outlook:

        Once you have validated your idea and learned what your customers really need, you can automate the process or hire people to do all the grunt work for you.

Recommended Posts