Building great data products
On why data products don't give out-of-the-box value to teams - People need to unlock them FROM the product.
Human beings remember stories and I use that style to narrate. If you want to read tech blogs on the subject, that is best articulated somewhere else. I can throw perspective on the right brain nuances of building - alteast in this post.
A bulk of my professional experience has been building large scale distributed applications around Intelligence application (Search, Personalization, Autosuggest, Real time bidding, pacing & location based algorithms, Social graphs et al) and another one on the receiving end of running, evaluating product alternatives and partner conversations to co-build our roadmap with them.
Lately, I have also been involved in a few companies in my capacity as an angel investor - where I touch nuances of building the backbone of an institution via a framework of what I call - “People, Processes, Culture, Tools”. During all these conversations, I have tried redefine(for my own clarity) what we think of traditionally as a “boundary of a Data product”.
There are two main things we will cover in this post
Solving the right problem broadly for the “sector”
And continuing to solve the right problem for a specific customer for a very long amount of time, irrespective of changing dynamics for the customer
… which means creating hooks(that don’t drive you insane) to help unlock that value from the problem DESPITE the org nuances.
..but first all data products, answer the same questions in two stages
Stage 1: Build & validation stage
1/ What are you doing that is 10X better than any competitor? How? (Gold, if you are honest to yourself) There must be a way the customers are living with that problem - are they convinced that they wanted it solved YESTERDAY and not anytime in the future ?
2/ Are there a few of these customers with whom the PMF has been established? What is the core landing use case that is unaddressed by other tools?
3/ The problem of data that you may be addressing is very contextual (there will be upstream & downstream dependencies, attachment to pipelines & tools , for all the wrong reasons (What if it breaks - who will do the Pager duty !!) And all this deeply depends on the culture of the company (client). How will you solve for this with velocity that is decoupled from these realities?
Stage 2: Early Mocks & proof of concept stage
So, before you even start your fancy little Data platform toy - understand these questions before you start the “let’s get to work together”
1/ Who approached who ? Were they actively looking for something in your category ? Or did you push it down their throat ? Or bumped in at a conference / vice versa. It’s OK if you initiated this, but what parts if any amused them positively. You will be surprised to know that a lot of Organizations do POC’s because some engineers want to stroke that “We did a proof of concept and tried something new” bug - try to gauge this early on and get a real sponsor on it. Avoid the cognitive bias of interest for intent.
2/ What is the distribution of power - Who is following up with whom - Oh, Man !! this one alone is Gold and is a easy give away on the alignment of incentives
3/ Continue to validate and revalidate when you see a loss in momentum. When you know, well, you know :)
And now… shifting gears ..
Stage 3: Mass adoption : Understand the context - People, Processes, Org Maturity, upstream & downstream product stack
There are a lot more questions in each dimension - the larger goal is to help you not just look at your product in vacuum, but also the context in which it is used by your customers and how it impacts their lifes, dreams, fear & goals.
Stage 4: Continual user delight : On why Products often don't give continual value to teams after the stimulus is removed
Product doesn’t self present its value to people - People unlock value FROM the product
But, first a small detour - Ever been to a marathon ? How does that experience feel different even if you are not the one running ? The part where every one feels is on the same team and experience a collective moment together.
Great products try to mimic this feeling of rush, communal togetherness, joy, serendipity, stoke curiosity all wrapped in one. Think of the erstwhile version products such as Facebook, Linkedin, Slack - and then something happens. The users stop experiencing the newness, or start experiencing a new range of problems that are a pure side effect of this new product !!
Eg : slack that was meant to replace email changed the way we communicate forever. In some cases, our communication became more chatty, more interruptive, fragments across channels & threads all those unread mentions that frankly wouldn't have gotten created because human beings just cant/wont type as many emails as they do text messages.
Well why ?
Because, the cost of writing the full context in a slack thread is higher, where speed / response is the goal. While in email - Thoughtfulness is the goal.
You can see the impact of a different product trying to achieve the exact same goal = (get the fk’ing work done!!) - has one people ? And hence, new areas of user problems will emerge at this inflexion point. Because at this cusp new habits are formed, and a passive distaste for a acquiring these “bad habits” has started to seed in the subconscious mind. At the cusp of this subconscious awareness - new products, services & verticals appear !!
Like in this case, when email(gmail) become a trash land “superhuman” appeared. SMS was replaced by What’sapp for the need of having multimedia sharing - an innovation that frankly should have been driven by telecommunication companies, but wasn’t. Whatsapp is now, becoming yet another forward junkyard. A whole new reputation based social media - that helps us choose our scarcest of the resources attention - will emerge. Products that are mindful and not notification hoggers !!
The reason I give this example is to think of the cultural context in which your product functions. For the rest of this post I will assume you are building a data product (Because well, that’s my strength) - and that there are upstream and downstream dependencies & cultural context that your product needs to understand.
Truth is most organisations have products that are much like public goods - in that the functional team that endorses it is the one that benefits the most of it. The tragedy of commons though is that, this benefit is inversely proportional to the actual cost it must pay to get the adoption - in that the network effects unlock the full value of the adoption, thereby justifying the cost.
Imagine you are building a query execution engine that replaces a Cloud offering (Think Big Query, Redshift etc ) - imagine you propose that you will evict all and save cost and improve SLA’s , and you know your product can definitely deliver that. But you also anticipate that to UNLOCK this value there will be a massive disruption in the customer’s daily workloads.
In this example then, you are sitting at a much downstream - whatever you do will effect all upstream usage & disruptions. Having these answers will take you to Stage 5
Stage 5: Stuck between POC & large scale customer love ? Do things that don’t scale : White glove product service
💡💡 Every functional team has a few people who will go to war for YOU. Your goal is to find this perpertually unsatisfied people. And the way you find them is to find a set of usecases, provide out of band white glove, high touch support - to build usecases where other products won’t. As a startup you are always defering your own mortality rate , in early year, or even beyond growing out of teens & maturity. This framework lets you do just that.
Another small (personal) detour here. Quite recently, a sizeable portion of my financial assets was deployed with a wealth fund manager. I was unhappy with them with no evidence of compelling returns, even after discounting for Business cycles.
So, I pulled the plug one day from this partner, and created my own White glove service. I shortlisted a few wealth fund managers - called them and proposed that I would deploy a portion of this amount with them - but they would have to help me exit these clunky positions. To test how well were their incentive aligned - I had my own private asymmetric information on performance of these assets. When they did get back - this made the decision effortless !! This I wish was a whiteglove service that all financial wealth managers deployed.
And now, apply what would a white glove service look like in your customer’s world.
Need a proof point : Gartner predicts that soon “89 percent of businesses will compete mostly on customer experience”, but - “while expectations for the importance of CX are rising, CX budgets are not.” (2018) - this also opens up the need to develop what I call as frameworks to provide a higher quality of overall product differentiation through an integrated & strategic white glove offering.
This combined with a superior product will create massive non-evictable MOATS for new age products and have deeper wedge in the mind of the users.
Likewise in the data platform example above, a great team should realise the use case of the middleware layer in between in a reasonably mature org (since they will have a lot of the suboptimal & fast growing “tool jungle”) and try to discover a way to service the user needs around those, with your product. The important point to note is that - from your lens as the partner, you would still need to map the value of this to your roadmap and the cost of this white glove service as seen from lens of the customer you are trying to service. But the best part is here - you don’t have to solve these problems right away - but somewhere in a certain future that establishes a committed intent and gets the incentives aligned, like your customers have never before been.
Invent a packaged solution - a lego blocks of sorts that no one else can - that clearly has incentives of all these players, clearly aligned AND also incredibly painfully communicated. And only then, you will find you 10x pod of gold.
Story Over !! finally, the gold : Nothing beats product clarity(or lack thereof)
In another world, I whetted a product roadmap for VC for an early stage startup and this was one one of the response I received for my review
... so we had a call with the team today. There was no real product clarity and I think they haven't even thought that through. Your instinct was spot-on in terms of product focus (or lack thereof). We have passed this for now and I have given them the feedback that they have to focus on sharpening their TG. <...Some personal context, withheld>
Honestly, this happens so much that we forget that all products and esp data products solve real problems for real users. Data products are clunky, dry & platformish, with a lot of mis-selling & demo wares, which is why great teams need to make the user FEEL having already used them and their most burning pains go away.
The more that experience is real & vivid - the better that experience of the joy in the pain of the marathon !!
Coming up !!
In the 2nd post I will breakup the data platforms product strategy and how to have a multi year vision with the immediate building blocks. The recorded talk is here at DevCon 2022
