Getting hypotheses to test [Datahub Cloud]
Getting hypotheses to test [Datahub Cloud]
2024-05-03 Rufus + Daniela DH Cloud longer term vision
Agenda
- Set the objective for today - to come up with the longer term vision
- Review current state
- Sketch Current user journey https://app.excalidraw.com/l/9u8crB2ZmUo/2Yi8iB94NSu
- Generate actual hypotheses and select ones that we want to test https://docs.google.com/spreadsheets/d/1kauprqTV0uh6y8nDagiDoFdnc_iUMzgFYKaIry20myc/edit#gid=0
- Create a hypothesis tree https://coggle.it/diagram/ZjTjjB8BY6WpL076/t/we-can-build-a-datahub-stack-based-that-generates-10k-mrr
- The value of being on reddit, hn (good for announces), stackoverflow, dev.to (see ### Marketing ideas)
Current game
Our game is to get to at least 10k of recurring monthly revenue
eg. 2000 users paying 2 a month
Current offering
We have this thesis about the power of markdown // Markdown publishing platform with extensions
On the landing page, we've got 4-5 ideas on what we can do. We've got a product that fulfills all these ideas with some little tweaks.
Instead of guessing in the dark, let's build a landing page and test.
New landing pages
Test the beginning of the funnel by testing the landing pages and the interest they generate.
The issue we are getting on datahub is that we get a lot of people who aren't necessarily interested in the product
PKM Datasets
////
Markdownit.com
Prequalify - some money $$
Marketing ideas from Rufus
Put yourself in the shoes of someone looking for tis product? Where do they go? What do they search for on google? Where does that take them?
Often to reddit, stackoverflow etc?
- Example: I rufus search for "markdown wiki" and i end up on reddit with a product pitch https://www.reddit.com/r/selfhosted/comments/10cmfvo/wikmd_a_filebased_wiki_in_markdown/
- We can do this! We can pitch our products
- Plus lots of opportunity to learn, post on opportunities etc
- Post on reddit
- you can get feedback
- you can draw attention to the product
- hack n use (good for announces)
- Post on dev.to
- Post on twitter
- Post on stackoverflow e.g. https://stackoverflow.com/questions/30600431/is-there-a-way-to-create-a-simple-markdown-wiki-with-github-pages?newreg=e810cf8bda80482a9b85c03da5cf515e
Daniela NTS valuable features hypotheses
- There will be real value in adding search functionality WITHIN the site itself + within the whole db of sites (coming back a bit to datahub as the data marketplace)
- I can sense that there is some confusion in some users from what datahub used to be and what it offers today..
- There will be real value in adding catalog functionality and making it great so that I can organise my datasets and make them more discoverable // or have collections and grouping of my datasets
- There will be real value in having some overview of what there is on my page ("On this page")(Flowershow has this already)
- There will be real value in having a button / option to navigate to the Github repository from the published site
- There will be real value in having an option to navigate back to the user dashboard from my site and be able to click "edit my page"
- Having to Sync is annoying if there is more than one person contributing to the site (which is one of the value propositions)
Datahub to remain about publishing datasets and searching for datasets
.. "Social media" for data engineers/data analysts?
Hypotheses
For each hypothesis β¦
- A testable claim β¦
- Why we believe that β¦
- How much value do we get from knowing that? How it contributes to the product going forward
- Is it testable? How expensive is it to test?
π‘ 5 Solutions and promoting them
We have 5 solutions:
- π₯ Dataset publishing
- Data story publishing
- Docs publishing - ? flowershow
- π₯ PKM publishing - currently flowershow
- Simple website from markdown
Create a landing page for each - can be own mini site, build with DataHub cloud.
So I think we want to have a solutions page for each of the major ways you can use DataHub. For example, you can use it to publish docs, you can use it to publish a personal knowledge management system, you can use it to publish a dataset and you can use it to publish a data story.
And then, even right now, we want to go and advertise that solution to the relevant audience if we can. There is a risk, of course, that we're spreading ourselves across several audiences but that's something maybe worth it to see what we get traction with the same tool. We can test even just building a very simple landing page for each of them and see if we can drive users to that place.
2024-04-26 [Daniela and Rufus] DH Cloud hypothesis
Our hypothesis about what a funnel looks like
Hypothesis 1
- Our waitlist conversion is a bit low.
- We think waitlist conversion should be ~10%
- https://www.reddit.com/r/SaaS/comments/14bw84g/whats_your_opinion_on_waitlists/ this thread suggests < 5% conversion from waitlist to product trial
- https://news.ycombinator.com/item?id=32276057 - not useful in terms of info
- https://www.indiehackers.com/post/whats-a-good-signup-rate-on-an-idea-validation-landing-page-15aa75057b
Hey! I'm the founder of LaunchList (https://getlaunchlist.com) - a tool that helps manage pre-launch waitlists for products.
I recently conducted a survey with a selected group of users, on average, the waitlist to product registration rate falls between 5-8%. But, there were some users who achieved a conversion rate of over 20%!
The rate of conversion depends on how quickly you can develop your product after receiving validation.
- Our waitlist conversion is X opens, Y calls, Z trials, A paid customers
- We think waitlist conversion should be ~10%
- Wrong type of people are discovering (?? we are reasonably clear on the website)
What's our hypothesis about why the conversion rate is low
- π₯ Sign up and then a delay => immediate invite to a call or a trial (how can we test this)
- β People don't want a call, they just want to use it => give them direct signup link
- β People don't really know what this is, they just signed up => send them some screenshots / short video along with invitation to trial
Hypothesis 2 after reviewing the waitlist
We have a quite junky sign-ups for Datahub Cloud
Add to the form: What would you be interested to pay�
TO TRY OUT
- Immediate offer to use the app (split this 50/50 for e.g. 100 people recently on waitlist)
- Immediate offer to jump on a call with me (this we run for all new signups in next week)
Immediate access | Get on a call immediately | |
---|---|---|
Most recent 100 signups | 50% | 50% |
Next week of signups | 0% | 100%% |
Extras
- People think we're "just another SaaS company" => do something out of the ordinary in our sign up email I think we could mix this in to the previous item
- People think they are going to get pitched/hard sold on f2f meeting => be up front. say product costs X, we want to set them up β¦ (less likely to work here b/c i don't think our main page does enough work to really show what the product is)
- Getting unqualified leads in the funnel => put a video and more info on site so people who sign up are more aligned.
- People are not convinced of the value? => send through some cool examples of what they could do
Market-fit Hypothesis
General context:
- Looking at Vercel logs, we have 28 signed-up users
- 19 Datopians
- 6 from personal network (we contacted and asked them to sign up and test
- 2 from the waitlist
- 1 from Bizdev
Clearly, the problem is not that the people are not finding the product valuable or that they are encountering technical issues. The problem is BEFORE the sign-up process comes into play, ie. it lies in the marketing/sales funnel.
We have no actual users yet becauseβ¦
- ππ₯π₯ Not enough people know about it (lack of marketing and/or visibility) - Quite likely although we have exposed it to some people (eg 20-30) and nobody is actively using it
- ππ₯ Not the right people know about it (targeting the wrong audience) - Very likely because we don't know exactly what is the "right" audience atm
- ππ₯The right people know about it but they don't understand it. Ie. Not quite clear what the product does and who it speaks to (unclear messaging or value proposition) - This is also likely because we aren't being super clear + there may be a confusion around the old datahub versions
- π There is no need for this product on the market (lack of market demand) - Less likely
- πThe onboarding experience is difficult and "costly" in terms of energy and time needed (poor onboarding experience) - Less likely. Yes, it requires showing up on a call and talking to a real human being but if people were motivated this shouldn't be an issue
- π The product is not valuable enough (insufficient value provided) - Less likely because they haven't even tested it
- π There are similar apps in the market that are more established or offer better features (competitive landscape) - Less likely because they don't even know all the features before they have used it
- π It's not clear that it's free to use..? (pricing concerns) - Maybe but not the core problem for sure
- π There are bugs and performance issues => users abandon it before fully exploring its capabilities (technical issues) - Less likely because that would've been the case if the users were signing up and then disappearing. They are not signing up in the first place
- π Users don't know anybody who is using it and benefiting from it (lack of social proof?) - Increases the chances of enrolling people but not the core problem
=> Conclusion: The primary challenge lies in the initial stages of the marketing and sales funnel.
- While there is likely market demand and inherent value in the product, the core issue is the lack of awareness and clarity around the product.
- Potential users are not signing up because they are either not aware of the product's existence, or they do not fully understand its value proposition and relevance to their needs.
- What is missing / what we should FOCUS on:
- Define the target audience (TBD after the product hypothesis gets refined)
- Refine the messaging and increase the visibility
Product hypothesis
What use cases can we think of..? DH Cloud is the go-to platform forβ¦
- Publishing datasets **This is the focus currently but it may require some more
- Publishing data-driven stories **This is the focus currently
- π Creating simple data portals Not a fit. The enterprise version is already simple and much better
- PKB: Publishing personal notes
- π Publishing academic research docs Maybe but we'd need to compete with eg. ResearchGate, Google Scholar, Academia.edu, arXiv, etc.
- π€·β Publishing technical and/or project documentation - Maybe but the data previews and data viz features won't be utilised this way which would be a pity => this could be a side use case and not part of core messaging
- π Creating blogs - Not impossible but too many competitors that are offering much easier and better blogging experience
- π Publishing educational resources - Not impossible but the current features won't be very utilised
- Creating some eg. performance reports - Possible but we need a more specific use case and it may require integrations with other plattforms
- π Creating event pages? - Not impossible but the current features won't be very utilised
Questions
- Why are the waitlist subscribers not
[Daniela] 25 Apr DH Cloud backlog shaping
Situation
- We have a usable product that is working
- We have an assigned developer working on improving the product, fixing bugs, and adding new features.
- We have 7 open/pending issues in our Epic https://github.com/datopian/datahub-next/issues/330
- We have other ideas and backlog items in our inbox https://github.com/datopian/datahub-next/issues/246
- We have 41 open issues in datahub-next repo https://github.com/datopian/datahub-next/issues/
Problem
- We don't have a long-term vision of the product
- No sense of the direction we should take
Rufus: i read this as: We don't have a clear hypothesis to test atm.
Solution
- We need actual users with actual feedback so we know what to prioritize and how to shape the product vision going fwd
- We need to play with the product more ourselves: we will use this β¦
- We need more rapid testing and more aggressive onboarding of the waitlist subscribers
=> Daniela will focus on two things:
- playing with the product daily. Working on her glucose revolution repository to start with and then expanding
- onboarding WAITLIST subscribers. Analysing and prioritising the feedback (ours and peoples') and translating it into issues for Ola to work on
Context
- The issue atm is we don't have a really clear hypothesis we are testing
- We aren't clear about product-market fit
- We certainly haven't found it atm
- But we don't know if that is b/c a) product is wrong b) not enough people know about it c) people who show interest aren't being onboarded properly d) something else β¦
Option 1 tests the the hypothesis
- We would use this product for something and find it useful
Option 2 i think would help with some refinement e.g.
- either create clear separate landing pages for specific products and see if they get interest
- work on some specific part of the process e.g. in the waitlist subscribers