[Follow along] 7 Days from Idea to Product (37)
Time to launch a little experimental follow along.
Ship a new product in 7 Days*.
From basic idea to working product.
The move by STM announced this week is pretty exciting.
I guess that you may be thinking about launching your own product?
If so, this follow along is for you.
Hopefully it will give you some insight into what may be involved.
I will write a daily update right here in this post on STM.
Make sure you subscribe ^ up there click on Thread Tools ^.
25 September Update
We launched last week - Ensider.io Traffic Insights.
If your new to this thread - start with the post on Medium first.
If your looking for the latest update - scroll to the bottom.
I will post the daily update as replies to this thread from today.
* The small print - I avoid working at weekends and I dont expect the people I work with to do so either. So when I say "Days" I mean normal business days. Again feel free to call me a cheat. Dont care.
Will be following this one as I am becoming more and more certain that I need to develop a product of my own soon.
Subscribed!
I love this kind of follow alongs!
Hope you will post also some of the technical challenges that you will go through the development.
Subscribed as well!
Thanks for all the interest guys. We're making good progress. Proof of Concept completed and we can move full steam ahead. I have updated the main post at the start with a summary of yesterdays (friday) activity.
Originally Posted by erikgyepes
Subscribed
Subscribed!!
Wow, talk about speed to market!
This will be pretty amazing.
So whats new? Can't wait to see some news 
Originally Posted by matuloo
I have posted todays priorities and will do full update tomorrow morning about how we got on today
Good day yesterday as the whole thing is taking shape nicely. Didnt manage to do every single thing we wanted to, but did some other things in stead. Here is the update. Ive added it to the first post in this thread.
Great Follow Along, and really love the lean startup principle, getting a minimum viable product to test the market quickly.
Fingers crossed for a success, but either way you're bound to pick up a few lessons along the way regardless.. Subscribed!
Yeah for sure @dazed1 - sofar we have identified 4 specific problems the market we are targeting is having that we are hoping to address with this product and are designing the plans in such a way so we can test which of those get traction.
Day 3 Update (13 Sep)
Focus: User interface
First off all - changing the approach because editing the full post here in STM was becoming a bit of a pain and too timeconsuming. So from today the full history will be here on Medium, and the daily updates will be posted as replies to this thread.
Priorities
- Add more of the core functionality into the UI
- Add user account data to our backend
- Setup basic waiting list -> signup flow and test
- Draft waiting list engagement emails
- Integrate intercom
- Setup spreadsheet to calculate pricing/plans
- Review options for analytics
Progress Update
Day 3 was primarily focussed on sorting out some key user interface elements and we made good progress here. The full dashboard is functional now including adding new items, viewing items and showing unread items.
We also implemented the full UI now for the login process. And added further metrics to the reports.
We also reviewed/considered how analytics is going to work. In the end it is down to Mixpanel and Yandex Metrica. Will run with both for MVP as our planning indicates we will remain within their free plans. Then after MVP we can decide which one suits us best.
There were a bunch of distractions on Day 3 as well which means a bunch of stuff I had to do did not get done. Gotta catch up on those on Day 4.
Challenges
Again no day is complete without challenges and complications.
We ran into our first issue when testing the waiting list solution (we are using Maitre App here). There was no easy way to track, remove or reissue submitted but unconfirmed emails in their interface. Lucily the guys at Maitre are on top of things and added this functionality that same evening
.The second issue is to do with our funnel. The landing page has a single clear action for the user visiting our page. When submitting the form, initially they will hit the waiting list. We will then promote a number of users to get access each week — but Maitre does not have a way of passing meta data in their Webhook. And it looks like they wont implement this until after we have launched.
So our options are: Implement a workaround ourselves, or Shift the funnel back to the original plan where the first action on the landing page is to join the waiting list.
Product/Market fit
The final complication was around questioning what the product is and what problems it is there to solve. It has multiple modules, and we reviewed whether at this stage we should split those out as separate products rather then bundling them.
The way we solved this was by going back over the problems we think our target market is having that this product helps them address.
In reviewing it we realised that at this stage we do not know which of those problems is the one important enough to our market that will drive them to put their hands in their pocket and pay us some money.
Therefore the best way to test this is keeping the product as a single product for now as part of the MVP and then use review and rapid iteration after launch to push emphasis on those areas that are most important for our market.
Day 4 priorities
Here are the priorities from Day 3, and the new priorities for today
- Intercom integration (Day 3 backlog)
- Waiting list engagement emails (Day 3 backlog)
- Spreadsheet to calculate plans/pricing (Day 3 backlog)
- Setup analytics accounts and integrate with site
- Integrate app with stripe and implement pricing plans
- Fix bugs in a key bit of functionality
Ill post the update about Day 4 as usual at the start of Day 5.
Day 4 - 14 Sep
Billing
Priorities
- Intercom integration (Day 3 backlog)
- Waiting list engagement emails (Day 3 backlog)
- Spreadsheet to calculate plans/pricing (Day 3 backlog)
- Setup analytics accounts and integrate with site
- Integrate app with Stripe and implement pricing plans
- Fix bugs in a key bit of functionality
Update
Posted this update later then usual mainly because normal life commitments take priority, even when doing something insane like building & launching a product in 7 days.
Despite the ☀️😎 hot weather here (30+ C) — we sorted a bunch of stuff out on Day 4 and caught up on most of the backlog.
Stripe is integrated from trial to upgrades. Just need to finalise pricing & plans and add some plumbing with webhooks and then we can charge customers 💳 when they are ready to sign up.
Pricing model also done together with basic cost analysis and rough financial model. Not sure we got this right yet at the middle/higher end. Need to ponder.
The third key thing we completed was the integration of Analytics.
We went for Segment to make that easy.
Segment allows you to work with a single JS snippet and API to manage all our tracking, analytics and support integrations. Then just switch on services you want to sent that data to.
It will push pageviews & events to Mixpanel & Yandex Metrica, and load the Intercom widget as well as sync all that data with our own Postgres database hosted at Compose.
This may make a good topic for a post in its own right. Let me know if your interested by responding here.
Besides that we completed the first draft of the email sequence and outlined the content (bullets). We also made some further UI improvements.
Challenges
Not many today.
The main issue with a key bit of functionality remains where we can’t seem to solve this nasty bug in our code.
👊 We’re not going to give in and remove it from scope right now.
Spend one more day trying to fix that bit. Hopefully we can do it!
D5: User experience (15 Sep)
Priorities for Day 5:
- Agree go live pricing plans
- Update Stripe and integrate required webhooks
- Join up Stripe with User accounts
- Integrate Maitre waitinglist webhooks with Intercom
- Implement key event tracking
- Fix or remove from scope the buggy functionality
Day 5 - 15 Sep
User experience
Update
A quick update as i’m typing this on Day 6 and still have stuff to finish so that we don’t need to work this weekend.
(note to self — don’t open that iphone7 box that just arrived)
We tweaked our pricing. I was factoring in build cost spread over 6 months. I’ve taken that out which enabled us to get down to better pricing levels.
All the plumbing for Stripe is also done with our App handling all the billing webhooks.
Ran into an issue trying to join up Maitre with Intercom.
The plan was to use Intercom to engage with people that submit their email for the waiting list. Turns out — intercom does not store a “Lead” until they send in a chat message.
Hopefully with some strong copy & automatic intercom popup we can get the user to do this.
Lastly our 2 developers managed to solve a nasty bug in a bit of key functionality late yesterday which means we can include it in the product on Launch! 🎉
Priorities for Day 6:
Pick name, get domain & setup SSL
Setup landing page
Plan out demo / walkthrough video
Test full funnel
Looks interesting.
I'm curious about this, is this a MVP or just a skeleton/base for a future product?
Shipping something is 7 days is either a very very simple product, a simple interface for an API or a productized service.
Originally Posted by crysper
Fridays update will be posted soon by the way.
@crysper - to be fair, if you cheat egregiously you can do an initial test for some quite complex SAAS products in 7 days.
Usually that involves using a Mechanical Turk approach, either in the classic sense (something that appears to be a machine that's actually human-powered underneath) or literally using Amazon's Mechanical Turk.
"Sprint" by Jake Knapp has a load of fascinating examples of testing stuff up to and including a hotel service robot in under a week by, essentially, cheating like hell.
Knapp's not quite talking about MVPs, but if you accept the definition of "Minimum Viable Product" doesn't necessarily include "can scale", you can produce some quite impressive MVPs this way and figure out the hard stuff once you know if you've got traction.
D6: Naming (16 Sep)
Should have posted this update on Saturday but didn’t touch my laptop at all that day apart from checking critical messages and issues.
So here I am, starting the final Day 7, writing this update for you.
Priorities for day 6
- Pick name, get domain & setup SSL
- Setup landing page
- Plan out demo / walkthrough video
- Test full funnel
Update
We picked a name and bought the domain . Throughout the last 7 days we kept a document with product name ideas. Choosing a name can be a distracting navel gazing exercise.
By timeboxing 5–10 minutes brainstorming each day, you stop yourself from falling down this rabbit hole.
Then on Day 6 we reviewed the list, checked which ones were available, picked our top 3, and decided in about 20 minutes on the final domain.
Oh … you want to know what the name is?
You will have to wait till this afternoon when I will post it here.
Changing domains did cause and issue with auth0.com because the domain you setup for the account & callbacks can’t be changed. So we had to setup a new account figure out why certain things were not working like before.
Following that we used Cloudflare to setup DNS because it gives us easy SSL.
Why did we not choose Let’s Encrypt?
Let’s Encrypt certs are used for lots of grey/black services which means your domains are more likely to get flagged when using Let’s Encrypt.
We setup the landing page base html structure with some dummy content.
We used the upside-down landing page format as a way of structuring the page. For theming we keep it simple — just vanilla Bootstrap with some custom CSS.
Finally we tried to make progress with the funnel, specifically joining up the Maitre waiting list system with Intercom. We hit a couple of snags here:
- Segment does not support creation of Intercom Leads. So if you post Events to Intercom using segment, it will associate those with a User rather then a Lead. Its a known limitation of Segment. To workaround this we disabled Intercom in Segment when sending the Events we want to track, and using the Intercom Javascript API to send the events to Intercom.
- Intercom only creates Leads in 2 ways: 1) when the visitor sends in a chat message, and 2) through the Rest API. This makes it difficult to ensure that each person joining the Maitre waiting list also has an associated Intercom Lead for future engagement via Intercom.
We didn't manage to map out the demo/walkthrough video and will probably have to go live without it and add it in later.
D7: Final checks & launch (19 Sep)
So today is launch day. Lots of little bits still to do though. Here are the priorities:
- Final tweaking of Waiting List & Intercom
- Write copy for Landing Page
- Develop some API functionality to store some meta data about each Lead
- Full flow testing
- Implement some basic service/availability monitoring
Day 7 : Final checks & launch (19 Sep)
So day 7 was meant to be launch day. It wasn't. More info below.
Priorities
Lots of little bits still to do.
- Final tweaking of Waiting List & Intercom
- Write copy for Landing Page
- Develop some API functionality to store some meta data about each Lead
- Full flow testing
- Implement some basic service/availability monitoring
Update
So we did NOT manage to Launch.
Frustrating but a reality.
At about 10pm last night made the decision to call off Launch.
The core product & billing functionality is all ready & working. But earlier in the day we ran into some compatibility issues trying to join up different parts of the funnel.
Turns out — Intercom PHP SDK requires php5.6 and we were running default Digital Ocean Nginx+PHPFPM droplet with php5.5 which doesn’t have an easy way to update to 5.6.
So upgrading to 5.6 required us to add a custom ppa repo. First we usedppa
ndrej/php5–5.6 but 5.6 support in that was deprecated, so we switched to ppandrej/php. Not how I would like to do my installs but time was of the essence.Then of course the MongoDb drivers were not working anymore. Took us way to long to figure out why. (doh they were installed for 5.5).
Anyway — you get the idea.
The Lesson?
Earlier on the project be clear about how the different components of full solution join up, and make sure all required dependencies are in place and tested.
Remaining tasks for Day 8:
- Full flow testing
- Write copy for Landing Page
- Implement some basic service/availability monitoring
Final checks then soft launch here
D8: Bugs bugs bugs (20 Sep)
As I explained yesterday — we ran into some serious issues joining up the funnel on Day 7. This took us most of Day 8 to resolve. But we have done it and are ready.
We setup a basic up side down landing page, went through full flow testing (end to end test from initial website visit to funnel and signup). and implement NewRelic monitoring.
D9: Launch (21 Sep)
So here it is … well just after this Intercom service outage! We love Intercom and these don’t happen often. Fate i guess?
I present the result of 7 … (or is it 8… or 9) days of work by a team of 2.5 to take one of those “we should create a product out of this” ideas, from idea to a real product.
Ensider — Traffic Insights for You!
Here is an update from the last 12 hours or so. 16 on the waiting list. No promotion done so far so that is all just from the follow along post.
Hi tijn,
great follow along, i just registered for the waiting list.
Your testing domain 7days.imrat.com is visible on the lander maybe you should remove that from the screenshot.
thanks @gunnar - whats your email and i will bump you a spot to say thanks for pointing that out
Late Friday evening update
Some funnel stats first - no optimization at all right now as we are in early launch mode really just making sure everything works
This evening we took on the first 5 users from the waiting list.
Ill report back tomorrow how many of these actually logged in!
The flow is quit simple really -
We send them an email to confirm that they have an account.
They click a link to confirm, which will generate an email login link.
We are using Passwordless login at the moment - each login generates an email with a login link that expires if its not clicked.
Users can then see their dashboard which lists their reports
When you select a report - you first get an overall traffic overview with a quick screencap of the main homepage.
Below it is where the real value lies
Till tomorrow
Great work!
Thanks @cmdeal.
Todays update - nothing to update really - apart from:
* all 5 users that were given access have logged in at least once
* one user has logged in 5 times
* the welcome message we setup in Intercom had an error so was not sent - have updated this so i will be shown next time users log in
* we have no analytics in the app - need to implement that tomorrow
Most importantly - need to engage with the people on the waiting list, so will be preparing another report for them.
I will post a nice example of some traffic reverse engineering here as well tomorrow
Just one piece of feedback from what I have seen so far.
The domain name for the service seems pretty meh.
Since this is a consumer service and not an affiliate campaign, the domain address itself would much more important than for an affiliate lander.
Right now the domain ensider.io seems much too easy to mispell or get wrong.
Surely there must be some better .coms that are available? E.g. www.URLinsider.com
OK so basically its a new frontend of a similarweb api ?
Since I fail to see much being done about the actual product, the entire 7 days were spend on building the ui / billing / user flow / metrics ?
Great project btw, you have a very systematic approach.
@cmdeal -
Now question is - to swap or not to swap.
Share your thoughts here -> https://www.facebook.com/imrat/posts/10201966755487999
@offersdirect
Billing - Stripe integration pretty straight forward
Userflow - Just draw some nice pictures
Metrics - Copy and paste Segment code and enable some integration
Actually the bulk of development time went into:
* integrating with APIs to pull the data
* build a api caching layer
* job queue system
* landing page downloader system
Just sent out an update email to our waiting list. Decided to allocate 10 more spaces tomorrow and then to close. Our focus will shift to working with our subscribers to improve Ensider.
Last few days we've worked on the monthly traffic report email sequence so thats all set to go out to subscribers.
Posted this a few days back on facebook which gives you a bit of insight into what the traffic report is about:
Also doing some housekeeping this week - specifically moving code into version control and manage development, test & deployment properly.
Here are some updated metrics.
Total leads from waiting list: 58
Spots opened up: 10
Users logged in: 9
Average visits/logins per user this week: 4
Number of users upgraded to paid plan: 0
Waiting list engagement:
We sent 3 emails to the waiting list - update 3 sent just now. Here are the stats:
Complications
Couple of issues we are running into.
1 - I cant figure out how to do normal follow up sequence / drip emails with sendgrid. We can import the waiting list emails to sendgrid as a list fine, and setup a "Campaign" to send an email to that list. We can also segment those lists based on user data & actions. But how to F@ck do I send a drip campaign in Sendgrid, so that when we add an email to the list they get the first email, then the second etc etc.
2 - Intercom Leads - something is broken with intercom leads when importing the waiting list. It shows all 50+ leads added/updated, but then only about 50% of them show in Intercom. Support has been looking into this for me for a few days now but no resolution.
any updates on the development? i'm using the tool myself and think it deserves more attention since it's a very powerful tool if you know how to use it right
Tried signing up, but never received any emails.. is this still launching?
Originally Posted by nusolutionz
If we can implement one thing to improve it in 2017 what would that be ?
Originally Posted by bizfreak22
@tjin - I'd love to hear an update on the progress of Dev, growth hacking, etc. It's been really interesting so far.
Home > Vendors >