3 ms·

It is an opportunity, but not an easy one. Spreadsheets are good and fine until they're not (fine-grained permissions, collaboration, referential integrity and

It is an opportunity, but not an easy one. Spreadsheets are good and fine until they're not (fine-grained permissions, collaboration, referential integrity and validation enforcement, multiple data source aggregation and ingestion, easy visibility into history and auditing, features that Excel has but people don't want to learn, etc). You have to convince customers that what they have going is not actually working, which requires having a discussion about how just because poor data management practices have thus far failed to completely destroy your company, they are wasting tons of time and money. Many of them will have in the back of their mind that they can always go back to how things were before. So the behavioral change/change management aspect of a product that really does directly compete against Excel is very challenging.

stcredzero 7y ago
Spreadsheets are good and fine until they're not (fine-grained permissions, collaboration, referential integrity and validation enforcement, multiple data source aggregation and ingestion, easy visibility into history and auditing, features that Excel has but people don't want to learn, etc). All of the things you just mentioned could be handled using just two mechanisms: 1) An extremely low friction way of flowing data to/from spreadsheets 2) Automated publishing/versioning If a particular piece of automation for flowing data between and in/out of spreadsheets is the lowest friction for users, then that will be the way they use by default. Then that software will be able to provide collaboration, fine-grained permissions, referential integrity, validation enforcement, multiple data source aggregation, ingestion, and auditing. Tracking versions takes care of the rest. The trick isn't providing all of those facilities. We've known how to do that for many years. The trick is to provide all of those with near zero friction. What's needed is something like Dropbox, plus dataflow. Just let users save their files, and it gets versioned and uploaded into "staging" for the dataflow part without their having to do anything. Then, as people start running into issues of scaling, enforcement, etc, enable them to start incrementally adopting more administration features -- all while preserving the transparent "it just works" magic of their just saving the file.
la_barba 7y ago
It is extremely hard to convince someone they're doing it wrong, when what they're doing is working for them and is making them money. Its pretty much ingrained in our human nature, I think. Instead of telling them their solution blows, its better to gain trust by first addressing the pain points without any major rework, and then slowly change the system bit by bit, with each bit providing an instant usable benefit to the end user.
- henning 7y ago
  You're right, but in practice I don't see how it can be made that simple or painless. There is no way to switch from email and Excel to a third-party web application without major rework. That's what makes it hard.
  - la_barba 7y ago
    I think there are intermediate steps possible in some workflows. For e.g. A business app that accepts an excel file as input for further processing. So now you can let them keep the excel file with their existing macros/features, but give them additional features via this workflow.

stcredzero 7y ago

Spreadsheets are good and fine until they're not (fine-grained permissions, collaboration, referential integrity and validation enforcement, multiple data source aggregation and ingestion, easy visibility into history and auditing, features that Excel has but people don't want to learn, etc). All of the things you just mentioned could be handled using just two mechanisms: 1) An extremely low friction way of flowing data to/from spreadsheets 2) Automated publishing/versioning If a particular piece of automation for flowing data between and in/out of spreadsheets is the lowest friction for users, then that will be the way they use by default. Then that software will be able to provide collaboration, fine-grained permissions, referential integrity, validation enforcement, multiple data source aggregation, ingestion, and auditing. Tracking versions takes care of the rest. The trick isn't providing all of those facilities. We've known how to do that for many years. The trick is to provide all of those with near zero friction. What's needed is something like Dropbox, plus dataflow. Just let users save their files, and it gets versioned and uploaded into "staging" for the dataflow part without their having to do anything. Then, as people start running into issues of scaling, enforcement, etc, enable them to start incrementally adopting more administration features -- all while preserving the transparent "it just works" magic of their just saving the file.

la_barba 7y ago

It is extremely hard to convince someone they're doing it wrong, when what they're doing is working for them and is making them money. Its pretty much ingrained in our human nature, I think. Instead of telling them their solution blows, its better to gain trust by first addressing the pain points without any major rework, and then slowly change the system bit by bit, with each bit providing an instant usable benefit to the end user.

henning 7y ago
You're right, but in practice I don't see how it can be made that simple or painless. There is no way to switch from email and Excel to a third-party web application without major rework. That's what makes it hard.
- la_barba 7y ago
  I think there are intermediate steps possible in some workflows. For e.g. A business app that accepts an excel file as input for further processing. So now you can let them keep the excel file with their existing macros/features, but give them additional features via this workflow.