What feature should you build next? Here are some pointers.
Alex is the founder of Tanda, a payroll SaaS that went through the several stages of "product" strategy. He wrote a post on the things he learned as Tanda went from an idea to an established SaaS with several employee. Here's a TLDR version:
Phase 0: You're trying to validate the idea/product
Write what your product does in as a 2-sentence summary:
We are going to make an employee time clock that runs on an Android device. It will connect to a website in the cloud so you don’t need any dedicated IT support.
Alex says that interviewing something like 100 potential buyers sounds good in theory, but never followed correctly. This is much easier to begin with.
Phase 1: It's becoming obvious what to build next
When talking to a customer or prospect, if instead of asking “does it do X?”, they ask “how do you do X?”, and you can’t confidently answer “oh, you don’t do X, and here’s why”, then that is an obvious gap.
Here's an example from Tanda:
We didn’t have one on the app store for the first 4 years of Tanda. We got asked “how do employees install the app?” all the time. Our answer: there’s a mobile website , and everything important also happens over email and SMS5 . So you see - we’d explain to customers - you don’t need an app, but what you do need is a really accurate award interpreter, and only we have that.
Phase 2: It's becoming harder to realize what to build next
This is where many founders start to struggle. Especially if you have enterprise customers and they ask you to build something only they will use.
This lead Alex to the principle of prioritisation by MRR. Instead of asking themselves "f we build this feature we’ll win all these new customers” or “if we don’t build this feature we’ll lose all these customers”, they used MRR as a primary criteria.
Phase 3: Getting aligned
Tenda used OKRs across the entire company, and every 3 months the OKR would get refreshed.
If Alex had time machine..
He would:
a) Be less afraid of adding new products/markets.
b) Change/remove features if they aren’t working.
c) The best product strategy is to build what your customers say they’ll buy.
This is so amazing! This is exactly the reason why I started Pandalign.com to help entrepreneurs follow the customer driven product development with simple tooling.
hello
The idea with using MRT as the main criteria for deciding what to build next sounds great in theory but I am wondering how this works in practice? For example we have several channels of user feedback like in-app feedback, email, G2 and a yearly product survey. We find it super time consuming and challenging to map all that input to features. And furthermore we already saw the problem that even if people say they would use it, they still might not afterwards since it might not solve their problem fully. Anyone up for sharing their processes?
Mapping is a pain if you’re using Sheets or Airtable or similar. We use Savio.io to do this (full disclosure: we built it!)
If provides a way to log feedback from email, G2, surveys, etc with a Chrome Extension. That feedback gets linked to feature requests, and revenue data gets layered in from your CRM or database.
Which means you get a table of feature requests with number of votes and total MRR. And before building a feature you can email each requester to validate your solution.
We’ve been working on this after selling our last bootstrapped company, and have raised from TinySeed. Happy to nerd out on how to solve this (with Savio or not) anytime!
At my company, we use data analysis like word clouds to identify most-requested features and centralize feedback in Jira, and then validate with a small user group before full-scale development. This ensures we are building what's actually needed (most of the time...)
Let's say you've got a bunch of feedback for your email marketing tool. Comments include phrases like "better analytics," "automated follow-ups," and "third-party integrations." You feed all this text into a word cloud generator.
The words "analytics" and "automated" appear larger because they're mentioned more often. This visual cue immediately tells you that your user base is particularly interested in better analytics and automated features. So, you'd prioritize these in your development roadmap.
This reduces time in mapping all input to features. Customers often use similar keywords to describe problems. Thinking in terms of common problems, then figuring out the best solution, has helped me.
Thanks a lot! That's a good idea, will try that out.
Whatever you are building during validation stage should always focus on one thing which is getting user response about the product.
Think of it like glorified survey form where all the structure should point towards will people use it in long run, do people like this product and how is the product helping them.
You might pull some stunt like making the performance slow just to validate do people really need the product that even if it slow you will get some complaint etc .
Sometimes it not necessarily be a product I remembers pieter levels excel nomad list. He just use an excel which get a lot of attention and from that excel he make a website. There are some who write blog and from the amount of traffic the blog get they somehow know people interests. Some give stripe links asking for money first. Some request email just to check how much people will commit to the product. There is a lot of method to validate.
For me I am building an automated seo research tool for blogger and in the validation stage. If you are blogger and want to try out an automated seo research tool where this tool will suggests you the best topic and keywords that suitable for your business blog, checkout creativeblogtopic.com
is it a great idea to charge people in the validation stage? just to validate? my guess is that it would make it harder to make people want to try your product and hence provide feedback.
From the marketing point of view, you want to find your buyers asap. You also want to validate you’re reaching the types of people who will buy your product. However it seems that if you’re strategic about it you can do a freemium version to build a lead list basically and focus hard on converting a % of those. That is contrasted against promoting information or other things to create “traffic you own”
With my current product, a smart chatbot for Crisp Enum I have never had this problem.
My users always tell me what I should build. I review every single request and if I see the potential (which means this feature could be useful for many users, not just for this specific user), I queue or implement it immediately (depending on the demand level).
I believe this is not a problem at all - to find out properly what to build next.
But if your users are quiet and don't ask you to implement this and that, it may mean 2 things: