Building analytics software: Progress Thread

[MaximilianT]

Me and my partner @gilli after completing a coding bootcamp decided to move into the website analytics market because absolutely EVERY website needs to know their numbers and data.

We are currently building the script which then needs to be included on the website to be tracked.

We are already working on an MVP, however we are undecided on what the key value skews of our software should be. In our opinion, AI is an extremely important value skew, but we don't know if AI is enough as a value skew and USP. So our question: how many of you use analytics tools and see issues that bother you when using the software that can be done better?

We ask for your opinions, thank you!

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[MJ DeMarco]

website analytics market

You mean subscription tracking and analytics like Pebblebee?

I think that market is underserved and needs disruption. Companies like Zoho are garbage.

Or are you talking about a Google Analytics competitor?

 

[BizyDad]

Following

 

[MaximilianT]

You mean subscription tracking and analytics like Pebblebee?

I think that market is underserved and needs disruption. Companies like Zoho are garbage.

Or are you talking about a Google Analytics competitor?

I thought primarily about a google analytics competitor, but we are at the very beginning in the whole analytics space, so it is mostly still about solving a problem with a sufficient need in the whole big market.

Which analytics tools are you using @MJ DeMarco @BizyDad ?

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[MaximilianT]

@MJ DeMarco what exactly do mean with pebblebee.com ? If I search for it, I only find this site which is selling some sort of AirTags.

And what is the connection between pebblebee.com and zoho?

 

[gilli]

Week 2:
We did research and found a number of APIs and libraries that we can use to build analytics software along the lines of SimilarWeb. These include rapidapi.com, the REST api from Similarweb itself, and some libraries for charts and graphs like the one from chart.js.

We have also already managed to build a first small version of a dashboard, which we now want to fill with variable data and then equip with an AI (which we want for the MVP at least) so that it can interpret the data and support data-based decisions for marketing and sales by predicting the future.

The biggest difficulties we had this week were first finding the appropriate APIs and libraries and adapting the tech stack accordingly. The whole thing was complicated by limiting beliefs.

The goals for this week are:
Maximilian:
- Finish building the dashboard (equip it with more graphs).
- Make it variable so that you can generate data for different websites via inputfield and firestore database (currently the API is still stored with a non-customizable URL in Angular).
- Interpret the data of the dashboard with an API (e.g. with the one of GPT3) and generate answers to future questions accordingly.
- Pack the whole thing from a one-pager into a working user interface of an MVP

Gilli:
- Finish building the start page
- start pricing page
- build login-ui

 

[MaximilianT]

Week 3:
We filled the dashboard with variable data and found an API which we now want to connect to it to build some sort of "AI-automated report generating". At the same time, we need to make the dashboard look at least to some extend good and user friendly, right now it is just a input field plus some graphs and span elements.

The biggest difficulties for this week were the integration of the API because I worked never before with this marketplace and some design problems with the canvas.

The goals for this week are:
Maximilian:
- Interpret the data of the dashboard with an API (e.g. with the one of GPT3) and generate answers to future questions accordingly.
- Pack the whole thing from a one-pager into a working user interface of an MVP

Gilli:
- build some sort of blog page to demonstrate usability better

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[Tau Ceti]

I am interested in your progress.

I am actually a subscriber of Plausible but I also know of Fathom and also
Beam analytics

What's your USP and why should I go with your product instead of the established competition when the time comes?

PS: I am not trying to be an a##, it is a legitimate question. I want to learn more about your angle of attack.
Also I recommend for you to read the Plausible blog because it contains many golden nuggets in terms of marketing insights.

The founders(of Plausible) have been very open about their entrepreneurship journey.

PPS: those links are not affiliate links, I am not a shill, I just like Plausible.

 

[machinistguy]

In our opinion, AI is an extremely important value skew, but we don't know if AI is enough as a value skew and USP. So our question: how many of you use analytics tools and see issues that bother you when using the software that can be done better?

AI isn't a value skew. What do users care if you use AI or two rats behind a keyboard?

Sounds like you're creating a solution for a problem you haven't identified yet. Figure out who your customers are first (forums, corporate brands, dudes who just got a shopify store because they watched a dropshipping video, etc...) then reverse engineer your value skew.

 

[circleme]

I am actually a subscriber of Plausible

The founders(of Plausible) have been very open about their entrepreneurship journey.

Their story is absolutely incredible! It's one of the most open SaaS companies I know in the EU. I'm a big fan and didn't knew about them until know. Ty @Tau Ceti

Afaik(read) they have reached over 100k$ MRR a few months ago and I hope they will crush the 1M$ MRR in this year, as a lot of companies will have to act in the upcoming months as Google Analytics closes their doors for UA-properties and GA in general isn't GDPR compliant with a lot of EU countries and it seems they get more and more.

Any updates on your Chrome extension btw? I guess I have to start my own Execution thread as well, as you and many others have already shared such a big amount of useful information. A lot of things are going on on my end as well which I probably should share.

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[Tau Ceti]

Their story is absolutely incredible! It's one of the most open SaaS companies I know in the EU. I'm a big fan and didn't knew about them until know. Ty @Tau Ceti

Afaik(read) they have reached over 100k$ MRR a few months ago and I hope they will crush the 1M$ MRR in this year, as a lot of companies will have to act in the upcoming months as Google Analytics closes their doors for UA-properties and GA in general isn't GDPR compliant with a lot of EU countries and it seems they get more and more.

Any updates on your Chrome extension btw? I guess I have to start my own Execution thread as well, as you and many others have already shared such a big amount of useful information. A lot of things are going on on my end as well which I probably should share.

I don't want to pollute this thread with my updates so I ll just post on my progress thread instead.
Send your progress thread my way when you start it. I d like to read it.

 

[gilli]

They don't pollute anything, on the contrary they help us a lot. We are currently working through the Value skew of Plausible and are very amazed at what they have built. Can you please show me the link to your progress thread.

 

[MaximilianT]

I am interested in your progress.

I am actually a subscriber of Plausible but I also know of Fathom and also
Beam analytics

What's your USP and why should I go with your product instead of the established competition when the time comes?

PS: I am not trying to be an a##, it is a legitimate question. I want to learn more about your angle of attack.
Also I recommend for you to read the Plausible blog because it contains many golden nuggets in terms of marketing insights.

The founders(of Plausible) have been very open about their entrepreneurship journey.

PPS: those links are not affiliate links, I am not a shill, I just like Plausible.

Thank you very much! Plausible is great, @gilli and I are working just now on their value skews and blogs. Our main USP would be the AI Integration in a similar way as Tellius is doing it. A dashboard with good data is fine, but for data-driven decisions numbers are just not enough right? There has to be some sort of "digital data analyst" who assists the actual analyst and partly or even fully automates his role sooner or later.

The reason, why you should choose us instead of the established competition is the main thing @gilli and I are working on just now. Our plan until now was to put many small value skews like much faster site loading times than Google Analytics, open source code, 2 click import of the data from the former software, etc. together and combine them with the AI mentioned above.

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[gilli]

They don't pollute anything, on the contrary they help us a lot. We are currently working through the Value skew of Plausible and are very amazed at what they have built. Can you please show me the link to your progress thread.

 

[Tau Ceti]

They don't pollute anything, on the contrary they help us a lot. We are currently working through the Value skew of Plausible and are very amazed at what they have built. Can you please show me the link to your progress thread.

Sure thing: EXECUTION - Building a browser extension empire

 

[gilli]

Thank you very much! Plausible is great, @gilli and I are working just now on their value skews and blogs. Our main USP would be the AI Integration in a similar way as Tellius is doing it. A dashboard with good data is fine, but for data-driven decisions numbers are just not enough right? There has to be some sort of "digital data analyst" who assists the actual analyst and partly or even fully automates his role sooner or later.

The reason, why you should choose us instead of the established competition is the main thing @gilli and I are working on just now. Our plan until now was to put many small value skews like much faster site loading times than Google Analytics, open source code, 2 click import of the data from the former software, etc. together and combine them with the AI mentioned above.

And after plausible we will work on fathom.com and beam analytics

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[Serito]

I used Matomo for a couple of personal projects, i found it easy to install and set up (it comes with a tag manager too, option to disabile cookies, etc.), You should give it a try.

Are you planning to offer an on-premise solution for your software?

 

[MaximilianT]

I used Matomo for a couple of personal projects, i found it easy to install and set up (it comes with a tag manager too, option to disabile cookies, etc.), You should give it a try.

Are you planning to offer an on-premise solution for your software?

We'll check it out, thank you. Regarding on-premise solutions, we won't offer such a model, we will offer it as a web application with a monthly or annual subscription cost.

 

[MaximilianT]

Week 5:
The feedback of @Tau Ceti was extremely valuable for us, especially this simple question:

What's your USP and why should I go with your product instead of the established competition when the time comes?

and his suggestions to check out plausible, fathom and beam analytics. We spent the whole last week on reading the blog posts of plausible and fathom and are now on a much better way to have USPs compared to plausible and fathom.

We made a huge table with USPs compared to plausible too. It was pretty obvious that Uku and Marko have pretty much copied fathom and for example the about us page on plausible looks nearly identical as the one from plausible.

As a result of the whole restructuring of our strategy, we had to rebuild the tracker with the knowledge from the scripts from plausible and fathom and while I was working on that, @gilli worked on the site and mostly on the blogs.

The biggest difficulties this week were for @MaximilianT coding problems and the new synthax and coding style of certain scripts and for @gilli problems in css and UX.

Funny problem we encountered the week before last week: Italy's government in Rome banned ChatGPT because of security concerns with the AI itself and because of data security and privacy problems and we literally just needed 5 minutes to find a working VPN service for chrome – after that, our computers thought we were in the Netherlands, so no more problem with stupid italian laws

The goals for this week are:
Maximilian:
- implement all our value skews from the table we made regarding the tracking script
- pack the data collected with the script in the dashboard
- if all of this is completed: build the login

Gilli:
- write the single blog posts and make it all variable

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[Tau Ceti]

I am going to be honest, I am still not seeing the USP here.

The AI thing doesn't really appeal to me.

You said that the AI will be able to answer all the questions regarding my site's performance.
The thing is that a simple dashboard can already do that. I don't need to ask questions to an AI because everything is available at a glance.

What I am expecting from a good analytics software:
- privacy(mine and the user's)
- no data mining/reselling
- light weight tracker ( should work on a slow connection)
- dashboard easy to navigate and customize if necessary

Granted, your idea may have changed since we last chatted in this thread, so this may not be relevant to you anymore.

Also once again, i want to put an emphasis that I may not be in your target market so please take whatever I say with a grain of salt.

You seem to have vision for your product which is great, so i don't want to put a damper on your aspirations.

 

[MaximilianT]

I am going to be honest, I am still not seeing the USP here.

The AI thing doesn't really appeal to me.

You said that the AI will be able to answer all the questions regarding my site's performance.
The thing is that a simple dashboard can already do that. I don't need to ask questions to an AI because everything is available at a glance.

What I am expecting from a good analytics software:
- privacy(mine and the user's)
- no data mining/reselling
- light weight tracker ( should work on a slow connection)
- dashboard easy to navigate and customize if necessary

Granted, your idea may have changed since we last chatted in this thread, so this may not be relevant to you anymore.

Also once again, i want to put an emphasis that I may not be in your target market so please take whatever I say with a grain of salt.

You seem to have vision for your product which is great, so i don't want to put a damper on your aspirations.

Yeah, AI itself isn‘t the USP, you are absolutely right. Therefore, the analytics software should / must be more than a analytics tool. Tellius is a good example for what we are planning to do. Of course you see all the CURRENT data and what numbers your site got, but does it also show, WHY these numbers are actually there? Which effort or strategy led to those results you have currently got? And how you can improve certain metrics over time? Plausible and Fathom aren‘t doing that and their reports do not contain any strategy of metrics to improve stats.

Their dashboard is what is – just a dashboard. It shows the numbers your website is currently producing, but it does not show clear reasons for problems and clear reasons for things that went well. So there is still needed some sort of data analyst person who reviews the data and finds the clear reasons for why things are working and why not. And our plan is to replace that needed person with an AI assistant, so everyone in the company can use the tool and directly gets what he needs and why that is so. The software should also create full scale marketing strategies and sales plans based on the current statistics of the site WITHOUT need of interference by humans.

What do you think of that?

 

[Tau Ceti]

Yeah, AI itself isn‘t the USP, you are absolutely right. Therefore, the analytics software should / must be more than a analytics tool. Tellius is a good example for what we are planning to do. Of course you see all the CURRENT data and what numbers your site got, but does it also show, WHY these numbers are actually there? Which effort or strategy led to those results you have currently got? And how you can improve certain metrics over time? Plausible and Fathom aren‘t doing that and their reports do not contain any strategy of metrics to improve stats.

Their dashboard is what is – just a dashboard. It shows the numbers your website is currently producing, but it does not show clear reasons for problems and clear reasons for things that went well. So there is still needed some sort of data analyst person who reviews the data and finds the clear reasons for why things are working and why not. And our plan is to replace that needed person with an AI assistant, so everyone in the company can use the tool and directly gets what he needs and why that is so. The software should also create full scale marketing strategies and sales plans based on the current statistics of the site WITHOUT need of interference by humans.

What do you think of that?

I actually would like to take some time to respond properly to you so I ll try to do that tonight.

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[Tau Ceti]

Actually I wont have time tonight so i ll just write this response quickly.

IMHO there is a problem with your current approach.

On one hand your vision is great but getting there will be hard an costly in terms of time and money.

You need to bring in revenue now. Not in 6 to 12 months which is what it will realistically take to do something like Tellius.

Customers don't buy features they buy solutions.
What is your solution?

You said that you want to create an analytics software but now you want it to do more. There is nothing wrong with that but those two things are not the same.

You are starting with the end vision in mind whereas you should be starting by getting something out there and see what your users tell you based on the feedback you receive.

To do so you need to get your analytics out there first, get some customers(paying not free ones) then build on top of what you have to reach your vision.

Which brings us back to the original USP question?
How do you convince someone like me to switch to your software right now? Not in 6 to 12 months when you have many more features.

You need to work by iterating on your idea.

Here is how I would do it:
- build your analytics software (just that , nothing else)
- get 20/30 paying users
- talk to your users
- build what they want
- get more users
- an so on...
- reach your product vision which may or may not look like Tellius

As it stands if you don't do this, you risk building for 12 months something that the market may not want nor need.

With the iterative approach you are doing market validation each step of the way, applying corrective measure everyday.

But for this to work , you need to deliver value from day 1.

So you need to convince the users of paid software analytics to switch now.

I am one of those users and as of today, I don't want to switch to your product because there is nothing to gain from it.

That is why I strongly urge you to spend some time to think about what makes you different from the competition today. Not in 6 months from now.

If your only edge is GTP-4 then I am sorry but that is not an edge.

Read my progress thread that i posted above about iterative approach to building a product. You'll understand what I mean.

 

[MaximilianT]

Actually I wont have time tonight so i ll just write this response quickly.

IMHO there is a problem with your current approach.

On one hand your vision is great but getting there will be hard an costly in terms of time and money.

You need to bring in revenue now. Not in 6 to 12 months which is what it will realistically take to do something like Tellius.

Customers don't buy features they buy solutions.
What is your solution?

You said that you want to create an analytics software but now you want it to do more. There is nothing wrong with that but those two things are not the same.

You are starting with the end vision in mind whereas you should be starting by getting something out there and see what your users tell you based on the feedback you receive.

To do so you need to get your analytics out there first, get some customers(paying not free ones) then build on top of what you have to reach your vision.

Which brings us back to the original USP question?
How do you convince someone like me to switch to your software right now? Not in 6 to 12 months when you have many more features.

You need to work by iterating on your idea.

Here is how I would do it:
- build your analytics software (just that , nothing else)
- get 20/30 paying users
- talk to your users
- build what they want
- get more users
- an so on...
- reach your product vision which may or may not look like Tellius

As it stands if you don't do this, you risk building for 12 months something that the market may not want nor need.

With the iterative approach you are doing market validation each step of the way, applying corrective measure everyday.

But for this to work , you need to deliver value from day 1.

So you need to convince the users of paid software analytics to switch now.

I am one of those users and as of today, I don't want to switch to your product because there is nothing to gain from it.

That is why I strongly urge you to spend some time to think about what makes you different from the competition today. Not in 6 months from now.

If your only edge is GTP-4 then I am sorry but that is not an edge.

Read my progress thread that i posted above about iterative approach to building a product. You'll understand what I mean.

You are completely right, we will work on the USP you have mentioned and we will execute the plan you mentioned (below).

Here is how I would do it:
- build your analytics software (just that , nothing else)
- get 20/30 paying users
- talk to your users
- build what they want
- get more users
- an so on...
- reach your product vision which may or may not look like Tellius

 

[Tau Ceti]

You are completely right, we will work on the USP you have mentioned and we will execute the plan you mentioned (below).

You don't have what I say to the letter. That's just one way to do this.

Not even necessarily the best way either.

My point is that many people have a vision of a product that does a 100 things and looks amazing and will be perfect and will be bug free and so on and you get my gist.

You can do that. Then in 12 months, you release your software and realize that actually the market doesn't need another Tellius or another analytics software.

But its too late, you have spent time and money and you have nothing to show for it.

Obviously there are exceptions to this rule as always. I know of a few guys who spent 6 to 8 months building something without talking to many users and released their product and it became a success but those guys are the outliers.

Whereas the approach outlined above reduces the risk of producing a dud.

If you cant convince 10 or 20 people to use your solution and to pay for it within the next 3 months, then you know that you are on the wrong track!

You can then pivot or you can abandon the project or you can make it better. Basically, you have options and you just saved yourself an extra 9 months of building and testing for nothing.

That means that each year instead of having 1 shot at releasing a product, you now have 4 shots. 4 chances to find your edge. 4 chances to make a product that people actually want.

In any case, don't take everything I say as gospel. I am just a barely successful app builder. I just started my journey and I do not have all the answers.

Please make sure to do your own diligence before deciding what to do next.

Also I apologize if i put damper on some things, this is meant to be helpful and avoid some costly mistakes that many first time founders make.

I know that, because I have made them.

With my latest venture, I do everything slowly. I build, I release, I listen. I adapt and change my plans. I try to release small features and then I wait for the feedback. I fix the bugs, and the I talk to users again.

I don't build stuff blindly and I don't invest a lot of time in anything unless I am certain that the market wants it.

Finally, there is a saying, if you are not ashamed of your product when you release it, then you released it too late.

Your product will look like s*hit when it first comes out. Its normal. It will break, there will be bugs.
All you got to do is fix one thing at a time. Improve little by little.

 

Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

[MaximilianT]

Week 6:

We completed the tracker and decided to make the node.js backend with the database work to know if it works before we would spend to much time on the frontend while nothing works on the server.

We want to keep it very small so we can build it faster. The same thing for the database. We'll use MongoDB because I already worked with it and @gilli is also familiar with firebase etc so it would not be a huge jump.

The goals for this week are:
Maximilian:
- complete the backend, which can post things to the database,
- set up the database
- create a first version of a working dashboard with the tracking script pushing data to the backend which then works together with the database

Gilli:
- write the blog posts
- set up the support center

 

[Tau Ceti]

Any progress on your business? Anything you learned that you want to share?