18
21 Comments

What are the most important metrics/analytics to gather and analyze in a SaaS MVP 🤔

Hi,

Working on a SaaS MVP at the moment. We are trying to keep everything lean, not just the product/code:

  • buyer persona definition
  • customer acquisition plan
  • product MVP

For the product, it's essential to gather data and analyze it to know where to put your focus.

❇️ What is for you the most important data or KPIs that need to be tracked at the very beginning?
❇️ What tools do you use to track them?

Thanks in advance!

  1. 7

    This is such a great question, JesĂşs! I feel like I have tracked every number under the sun and ping-ponged from tracking too many, to too few metrics over time.

    I think in reality, track as much as you can – it's helpful to know you have the data, especially in a raw form, going back as far as you can, because you never know what sort of queries you may need to perform in the future.

    BUT you don't need to pay attention to all of those metrics – ideally find ONE or at most a handful of metrics you care most about.

    In the early days, I feel one of the most valuable metrics to track is around Product-Market Fit. There is of course no "perfect" way to measure PMF, but I think the best (least bad) way to measure is by asking the question:

    “How would you feel if you could no longer use the product?”

    And measure the percentage of people who answer “very disappointed.”

    Supposedly, if you can get 40% or people answering "very disappointed" then it's a great indication of strong PMF.

    The folks at Superhuman wrote up a really solid explanation of how they did this, and it's also something we've been using at GoSquared with good results so far. Here's the post:

    https://review.firstround.com/how-superhuman-built-an-engine-to-find-product-market-fit

    Aside from this, as you build confidence on PMF, you may get clearer on a "North Star Metric" – the overall metric that best measures the value your customers receive from your product. This can be hugely valuable for aligning a team and ensuring everyone is focused on delivering value, not just driving revenue.

    Hope this helps, happy to chat more – this topic is hard to do justice in a single comment!

    1. 3

      Came here to share that resource from Superhuman as well :)

    2. 3

      Excellent advice. Came to say the same thing about product market fit.

      1. 1

        @hgld Thank you so much!

    3. 2

      Hi James, this is great advice, I appreciate it. Other than PMF, do you recommend tracking other essential KPIs related to the usage of the service such as retention, churn, etc?

      1. 3

        @jrbayguade Thank you Jesús! I think it all depends on the stage you're at – different things are important at different stages and it's easy to track more things. It's what to pay attention and actively focus on changing that's the hard part!

        I would certainly instrument metrics around feature usage (e.g. well considered event tracking on your most obvious features), and use these to help understand retention by signup cohort.

        For us, we recently went through an admittedly challenging task of defining our North Star metric (not as easy as it might sound!), and then chose just three "key drivers" of that metric (for us it was one around top of funnel, one around activation, and one around retention)

        Wrote about this a little more in a blog post about a growth course we recently did – if you're interested!

        1. 2

          thanks for the additional comments James, and the blog post was very insightful. I will definitely consider that course, although it's a bit early for us -we are just starting with the MVP and don't even have a landing page yet-

  2. 4

    I think the most important data to track for an MVP is feedback. What are your users saying about it? Where can it be improved? Anything else at this stage won't matter if you are not listening to your customers keenly to understand their problems and whether the MVP is actually a solution. There are a ton of feedback tracking tools but at this stage what you would need is not a tool but direct-to-consumer communication. Do things that don't scale.

    1. 3

      I would agree to this point. Getting people into a feedback questionnaire is crucial.

      • Email leads per week
        Great to keep track of the return on your outreach efforts. This is your most important step for getting product feedback. You should be sending these people a questionnaire and engaging as many of them as possible. Number of lead emails is your initial rate-limiting step.

      • Returning visitors
        Not so much a Metric, but observing a behaviour. Get in contact with them using your chat tool the next time they land. Why have they considered you enough to view your page multiple times.

    2. 2

      I find this useful as well, I don't think it's necessairly contradictory to other comments

  3. 3

    At the earliest stage, you should be tracking the priority of the problem your solving compared to the other problems your target customers face while doing the activity you're focused on (ie. the Stripe approach called Customer Problem Stack Ranking).

    Once you're clear that your value prop aligns with one of your customers' top priorities, then you want to move towards general metrics like: happy -vs- unhappy inbound messages, any referral that is happening organically, retention (MRR if paid, active usage if freemium).

    Imo, don't go overboard on quant at this early stage. Focus on understanding early adopter's experience qualitatively (ie. doing things that don't scale). Quant metrics are better for optimisation than direction setting...

    1. 2

      Thanks for sharing this article. I hadn't read it before or heard of the OpinionX product mentioned in the article.

      I also agree not to go overboard on quant as it doesn't provide the depth of insight needed at the MVP stage.

  4. 2

    Good question!

    What type of SaaS product is this exactly? It would help to know to give more specific examples.

    MOST IMPORTANT DATA IS QUALITATIVE AT THIS STAGE, BUT...
    At the MVP stage, the most valuable data will come from your customers, not from product usage data. I would focus on gathering qualitative data about user needs and their "jobs" if you have 10 customers or less. More on this below.

    You can learn so much more from less people directly, than you ever will measuring conversions and churn at this early stage.

    Metrics will only tell you the direction you're going in, not who is going with you on the journey and why.

    That said, and without knowing more about the type of SaaS product the key things to monitor over time are whether people want to use the product after learning about it, how much it cost you to get customers, and whether they stay:

    • Conversions (for one or more components of the product e.g signup form, free trials etc)

    • CAC (customer acquisition cost): its good business to know how much it costs to acquire customers. If you're using ads to attract people, pay attention to this.

    • Churn - how long before they leave? what percentage don't renew a monthly subscription?

    • Engagement: additionally, there may be some obvious aspect of engagement that could indicate the value of your SaaS to users. Without knowing more about your business, it's hard to say.

    Without talking to people, you will never know why any of these are good or bad.

    CUSTOMER NEEDS - USE THE MVP TO DO THE RESEARCH
    In the early stages, it helps to have people on boarded as "beta partners". Then you can potentially watch them using the product and find out what metrics are important to your users and to your business.

    You can either advertise that you're looking for beta partners and testers on your main website, or just reach out to people who sign up (if its not a breach of data privacy, look them up on linkedIn first). You're looking for two types of person: the "advocate" and the "generous sceptic" - easy to spot, the first will be super keen on what you're doing, but the sceptic will be more useful - providing honest criticism that is more useful to you but harder to hear.

    If possible, do some formal user testing. It doesn't have to be with customers to test some UX elements. If you can get customers or prospects though - its useful to test the same features in your product AND a competitor products. Ideally a product that your prospects use at the moment. You'll learn so much from testing other people's products as if they were your own and what users say about them. It's gold.

    TWO WAYS TO BUILD PRODUCTS:
    There are two ways to start building a product, either way direct customer research is a major component.

    1. do research first to understand customer "jobs" including their over-served and under-served needs.

    2. build a product first and then testing it and iterate on it (i.e. "lean" or "fail fast" - its the same).

    With (2), it is vital to realise you are actually doing (1) but using an MVP to attract people in order to learn about them. The most lean way to do this is not building a product at all, just a landing page with a signup form, then reaching out to people who subscribe and asking to talk to them. If possible and with permission, record the audio / video of those interviews on zoom or whatever. You will hear it all very differently days later. Even better, get an automated transcript of the interview (there are free services online for this).

    You are trying to understand how well your product meets customer needs. Given that you've built an MVP, I assume you:

    • have a hunch
    • spotted a gap in the market
    • are solving your own problem

    None of these justify using quantitative metrics over talking directly with users in my opinion.

    PERSONAS CAN BE A DISTRACTION
    If your buyer persona is a demographic / psychographic description about their habits and preferences, I'd throw it away for now but keep the original research (if any).

    Is the buyer also the customer and the user? Sometimes they are three different people!

    If you have less than 10 customers (or no customers) you're better off having real information about real people and their actual needs. Qualitative is better than quantitative early on. Personas aren't diverse enough at this stage, not even for buyers.

    FOR DISRUPTIVE INNOVATION, FOCUS ON JOBS TO BE DONE

    Quick overview; a JTBD:

    • Is solution-agnostic.
    • Results in progress when completed.
    • Is relatively stable across time.

    "help me store my documents online quicker than Dropbox instead of a usb stick" is not a JTBD statement. It is tied to existing solutions (dropbox) so there is only so far you'll be able to take your thinking with this.

    It is fine if you are happy to try to differentiate on brand but not disrupt the market. It tells you that the customer wants the same but faster. Dig deeper though, and you'll learn what job the person is actually trying to do.

    Better example: Keep my documents safe and organised

    No solutions in that statement and it is relatively stable over time. Jobs don't change, the underlying tech does. The job of listening to music on the go has changed very little since I had a Sony Walkman as a child, but the job i'm trying use it for has been stable over time.

    STANDARDISE THE NEEDS STATEMENTS
    Once you're clear on the 'job to be done' of the user, standardise the needs statements if possible so you can compare and rank them. There can be many, sometimes hundred of needs statements related to one job. It varies.

    There are multiple ways to frame 'needs statements':

    The Design Thinking way -https://www.ibm.com/design/thinking/page/toolkit/activity/needs-statements

    But I'm a big fan of Tony Ulwicks statement structure:

    (1) direction --> (2) Metric --> (3) Object of control --> (4)Context

    e.g. For a music app like Spotify, a need statement might be:

    Minimise the time it takes to get songs in the desired order for listening

    minimise = direction
    time it takes = metric
    get songs in order = object of control
    for listening = context

    People don't speak like this, but if you ask the right questions you can easily structure these statements afterwards.

    Later, you can survey a larger customer base to find out which statements are important vs how well the customer is currently satisfied with them. This will reveal under-served and over-served customer segments which work far better than personas in my opinion.

    Definitely look at Jobs to be Done theory. Read one of the JTBD books or Tony Ulwick's articles. Avoid the many inaccurate write-ups online. Here's a link:
    https://jobs-to-be-done.com/inventing-the-perfect-customer-need-statement-4fb7de6ba999

    WHAT DO YOU NEED TO KNOW FROM THE FIRST USERS?

    Frameworks aside, i think what you want to know from early users at this stage is:

    • where does this product fit into your workflow?

    • when would you use this? why/why not? how often?

    • what is the procurement process like at your company? e.g. how long is the sales cycle, who is involved, and what products are you trying to replace?

    • is this project used for a small "job" as part of a bigger "job" e.g. boiling a kettle to get hot water is a "job", but its part of the larger job or making a hot coffee.

    • in order to replace your current product, does ours need to be faster, better, cheaper or all three?

    This is longer than I intended, my apologies for giving you so much to read.

    I hope it helps :)

    1. 1

      Hey, thanks a lot for taking the time to write such a complete answer. I really appreciate that. I take good note of every piece of advice. Happy to see fellow Spaniards here btw :-)

      I also believe in the "jobs to be done" framework. I have listed the jobs, pains, and gains of the user by using the tool, and that leads to better know what value is being provided, and that can be tested qualitatively at every stage of the process.

      The product itself is somehow similar to Bugherd, but more focused on a specific niche. But we are literally still in a very early phase, nothing that can be tested yet, and not even a landing page yet. Working on all that atm.

      thanks!

      1. 2

        No worries, sometimes I get the urge to write (usually too much).

        To your point about pains and gains, they are an overlapping concept to JTBD in that they help to highlight if your product is meeting users' needs. There are subtle differences I notice in practice.

        Pains and gains can be mapped out with the Value Proposition Canvas (for anyone not familiar - it sounds like you are!), but JTBD done well/right goes a little deeper for me.

        Bugherd is new to me. Interesting product, landing page is clear with a strong value prop and audience segmentation.

        I look forward to watching your success unfold on your journey !

  5. 2

    At the very beginning, here is what's important:

    • Can you find and talk to lots of people who have the problem you're trying to solve?

    • Do they care enough about this problem, that they are finding work-arounds and inefficient ways to "solve" it today? Are they looking for better solutions?

    If you can do the above in a repeatable way, you've gotten a buyer persona and a customer acquisition plan. Then the challenge is to figure out if the solution you envision would actually solve their problem.

    I think it's all qualitative at first, you will know if you're getting these things done. Once you have an MVP that people are using in practice, then it's time for KPI's and spreadsheets.

    1. 1

      Thanks for the advice J! This was really useful and made an impact in my plans :-)

  6. 2

    I think this can/should be fairly minimal. Track each step of the funnel:

    • Landing page
    • Conversion form
    • Conversion

    Do this as page view events.

    You need to know:

    • How many people are coming in at the top of the funnel
    • What the drop-offs are

    That way you can see if your conversions are because of loads of traffic or because of a good conversion rate. And you can see how your marketing actions affect the picture.

    I recommend Amplitude. It's more manual than GA but you get only insights about things you care about. It's incredibly powerful, you can grow with it. And it's free.

    Also, I'd kind of recommend not tracking anything to start with, and adding stuff once you realise you need it. (When you wake up and think huh, I have no idea how conversion is changing over time - that's when it's most important to start tracking conversion.)

    1. 1

      I agree here... I'd say the first thing is to get customers. I believe we need to start by tracking the conversion rate and analyzing how discoverable the landing page is (SEO).

  7. 0

    As a (serial, compulsive) bootstrapper I like to track anything and everything I can see that will give me any impression of growth -- the inevitable killer for me is feeling like a product or business idea is stagnant. As soon as I feel Iike I've lost momentum the pushing for growth starts being a slog, and I have to fight to keep my focus.

    I actually built a tool that I use for myself that takes my current metrics (acquisition, payments, arppu, etc.) and forecasts them out for 12 months into the future. It's been a huge motivational boost for me to be able to see where a product is (hopefully!) going even if it's nowhere exciting today.

    Give the forecasting tool a look at https://saascast.io -- I really can't overstate how helpful it's been to me to be able to look 6-12 months in the future and say "yeah, this tiny side project is totally on track".

    1. 1

      this is cool, but we are not there yet. thanks

Trending on Indie Hackers
How I grew a side project to 100k Unique Visitors in 7 days with 0 audience 47 comments Competing with Product Hunt: a month later 33 comments Why do you hate marketing? 27 comments $15k revenues in <4 months as a solopreneur 14 comments Use Your Product 13 comments How I Launched FrontendEase 13 comments