- Getting the license key was a bit slow, I went through the process of getting a reminder before the original turned up (I guess I'm impatient - but it's annoying when I had to hunt around my Junk Mail/Deleted Items to see where the license had gone). Why bother with a license for a demo?
- I find the appearance is too different to Excel so is a bit jarring
- The animations are cute, but not sure that they actually aid usability
- Although my initial reaction wasn't that positive it did actually grow on me but I think it needs an obvious immediate impact to demonstrate a clear benefit over the standard Trace Precedents/Dependents
- I can actually think of one scenario where this would have been pretty useful. I was designing a industrial modelling application that had been "prototyped" as an extremely complex Excel application - I had to spend a lot of time showing people that the calculations in the spreadsheet weren't what people thought they were - so having some way of automatically generated documentation of parts of a spreadsheet would have been useful.
Overall - not bad, but I think you really need to nail the exact pain points this would be useful for. Maybe also drop the Tony Stark/Avengers example - it's cute but might not appeal to your target market.
Thanks for the feedback. The big pain point we are focussing on is when you are given someone else's spreadsheet that you are not familiar with. It then can take hours to get your head around how they have put it together.
In this regard, Slate is most useful as way of semi-automatically documenting the spreadsheet.
We made the Iron Man example specifically for HN - possibly not the best thing to pitch big financial firms with - fair point!
I've spent weeks trying to understand a large spreadsheet. Mind you, a lot of that was persuading people that I was correct when I said what it was actually doing was at odds with what they though it should have been doing!
I had some trouble reading nested IF functions, so I wrote a piece of code which translates them into flowcharts: http://excel2flow.asp2.cz/ (weekend project)
It briefly bugged out (dumping everything as a string into a single result block) on entering a ":" or ";" after the word "Error" in a string as part of the long formula I tested it with, but since then I've changed a few minor things and it's stopped happening :S
Thanks for testing! ";" is delimiter for function arguments in czech version of Excel (because "," is decimal delimiter). Parser tries to guess if you're using czech or english version and it probably guessed wrong.
Anyway, I wrote the parser by hand so I cannot guarantee it will work on every valid Excel formula. But I'm working on improved version which will use formal grammar.
First some superficial stuff: Installshield is not going to fly with the market you want to make real money off, i.e. big organisations who want to deploy through Active Directory. The Office addin installer can be put as a merge module into an MSI, that will make the overall install experience much slicker. The instructions on the web page that opens after installation say to look in tab 'Add ins', but it's in separate 'Slate' tab (Office 2013/Win8.1). The introduction tutorial is unintelligibly small on a high dpi screen (Lenovo Yoga Pro 2); so is everything else that doesn't scale with ctrl-mousewheel. I just aimed very well with the mouse to make it go away and I tried to figure it out on my own.
On to the functionality. I work with Excel models if not daily, then at least a few times a week. These are usually either models I get from partners and which I integrate with other models, or they are data analyses. Over the last few weeks I worked with an economic model and a demographic model I got from outside, and I did several analyses on land use/population data myself. They're always intricate and on the overall complexity scale (as far as Excel models go) range from 'medium' to 'high'. So I think I'm exactly your target customer.
I tried first with a medium-complexity spreadsheet (5 worksheets, couple of dozen 'variables' where the variables are usually 2d matrices and the analysis is in aggregating and disaggregating the input data in various ways). I tried Slate on several fields, but for none of them I could better understand what they were than by using Excel's build-in 'Trace precedents' or even by just reading the formulas. Slate shows only very little context; you have to scroll out far to see those 2 or 3 steps away, but at that point everything is too small to read. Furthermore, it doesn't really dissect formulas; for example, an 'if' with two VLOOKUPs is just a pink blob with a bunch of arrows coming in, without giving information on how or what. It also takes up half of the width of my screen, so there isn't a whole lot more I can see on the screen.
What is worse is that there is no description of each 'block' (could have taken the label in front or above the formulas in the original sheet); nor does there seem to be a way for me to enter them manually. For example, I can see the value in being able to annotate a 'box' (variable) in my model with 'Plot counts per district', then an out arrow with 'multiply with average amount of dwellings in low residential land uses for the transport zone', then the box where the arrow ends up with 'dwellings in low residential housing areas per district'. That doesn't seem possible.
When 'show detail' is on, some of the boxes have the upper left corner of some range of cells in them; I'm not sure how or why. They just seem to take up a lot of space.
It doesn't look like I can edit anything in the result. For example, I have a lot of 'if(iserror(formula), 0, formula)'. I don't want to see all of that - I just want to see 'formula'. The rest is an implementation detail.
When 'hide detail' is on, there doesn't seem to be a way to jump back to the Excel sheet.
I had a more complex sheet I wanted to try it on after this one, but even this medium complexity one was harder to understand (and I wrote it over the last few days, so it's still all in my mind) than the original Excel sheet is; so I didn't bother with the more complex one.
Sorry to rain on your parade, it must have been a lot of work. I don't see much added value though in its current form. I'm sure that with much polishing it can become useful. I think annotations should be first; then abstracting away implementation details of formulas, you have more ideas yourself probably. Slate also loses the spatial relation between e.g. columns next to each other; I'm not sure what to do about that, but it should make up for that somehow.
Lastly, the price is ridiculous at this stage. At first I though 'oh 50$ seems OK, then I saw '/month/user'. $50 one-off seems OK; I don't see who will pay 500 a year for this, especially with the basic functionality it has now.
Feel free to email me if you want to know more about the spreadsheet I tested it on. I'd love to be shown what I should look at that would give me more insight in my model; I didn't see it myself. A real-world example on your website (as opposed to an Iron Man made up one, which to be honest is quite juvenile) would maybe help, too.
Good luck with your product. Please don't be discouraged by this - I'm not trying to put you down, just honestly telling you what my experiences were, as you'll get very few reactions from people like me who decide to give up after 10 minutes; while those are exactly the ones you want to hear from to know what to work on.
From what you've said, I would argue that you're /not/ their target customer. I used to work with medium- to high-complexity Excel spreadsheets on an almost daily basis, and while understanding someone else's spreadsheet can be tricky in the first 3-6 months, after a while you just get used to working out where data is coming from, usually with the help of F2 (which highlights the cells which are contributing to the formula).
Which leads me on to the main point: 'Who exactly is the target customer here?'. If someone uses Excel so much that it makes sense to buy something to explain formulas, they're probably going to be good enough to make sense of formulas using Excel's in-built tools/just reading the formula.
Yeah maybe you're right, I just thought that from their price point they weren't targeting Nancy down the street who needs to decipher the spreadsheet with the proceeds of the bake sale. I'd consider myself 'entry-level power user', which I then by elimination thought of as the target market.
IFERROR was introduced in Excel 2007, and is not compatible with the old-school .xls files. I haven't had to worry about 2003 in the wild in a few years, but some tools may not have migrated to the new file format.
> is not compatible with the old-school .xls files
You can save as XLS (97-2003) files and Excel serializes the formula using the name `_xlfn.IFERROR`. The file obviously doesn't work in older versions of Excel but will work in 2007+
Oh that's very convenient, thanks. As another person replying said, I only migrated from 2003 a few months ago (don't ask) so I still have a lot of new quality of life features to discover :)
Sorry you did not find it helpful for the sheet you tried it on. We have not yet implemented a way of showing 2D nodes in an intuitive manner and this is one of our focuses for the next release.
You mention context - I think this presents possibly one of the biggest barriers to adoption. We included the ability to jump back to the cell in Excel as a result of similar feedback. There is definitely a lot more that can be done though.
You also allude to where we think Slate can be really powerful - documentation. If we were to implement the ability to annotate and group together nodes in Slate, you can give a general overview of how the spreadsheet is working, with the option to dive deeper and investigate the nitty gritty, if needs be. All too often you are given someone else's spreadsheet with no context, and this is where we think further value lies.
Would love to take this conversation to email to discuss further. I'm fraser@useslate.com
I want to contact you regarding another Excel addin idea. I could not find your email in your profile. How do I reach you? Alternatively, you can also email me at contact@solutionyogi.com
Great feedback, marred by a pointless "Iron Man reference was juvenile" comment. I for one found the example spreadsheet entertaining, and it made the product more appealing to me as a result.
Well as someone in their target group (or maybe not, as someone above remarked, I don't know) I did find it juvenile. What's wrong with saying that? Or do I need to carefully construct everything I say with 'this is just my opinion' etc - this is just an off the cuff report from a first time user to help these guys out, I don't think it's reasonable to expect I'm going to be all touchy feely.
IMHO, the concept is novel. Having worked in BI/data analysis fields, I have seen a fair share of large xls files with complex calculations. Can't really try now since blocked at work but will definately give it a shot on home pc.
It's a great idea, but I've been working in multiple enterprise excel hells and most of the places use macros, because the simple stuff just does not allow needed complexity.
What we have found is that one of the biggest problems in the BI space is the fact people are not able to code and resort to some of the horrific formulae we see. But agreed - including macros would add a huge new layer of complexity!
I much prefer reading Excel spreadsheet with complex formulae, than those with complex macros. It's much easier to follow things around and literally 'see' what's going on.
would be interesting to know what installed userbase libreoffice has. especially compared to excel. i rarely use spreadsheets, so libreoffice calc is perfect for me.
The add-in is currently bundled with all its dependencies so it will install quicker once you have downloaded it. Some of the bulkier files include the Chromium Embedded Framework dlls which we decided to include to get the slick animations.
Very minor detail, but numbers should be right aligned "tabular" numbers in your flowcharts. Digits like to be vertically aligned. I'm on Linux, so I don't know if that's just an artifact of the example on your homepage or if it's baked into the plugin too --- but that goes for the default settings in the plugin too.
seems like a really interesting concept, although I do think you might struggle getting that into the enterprise (or even medium/large business space) given the heavy installation.
Definitely going to see if it helps with me requirements of trying to unpick complex financial models built from spreadsheets.
Good point. We have had some issues with corporate security, especially in larger organisations, and are looking at alternatives. Moving to cloud software could be an option, but big financial firms are even less open to that that what we are currently building!
Im currently in consultancy for financial services and getting anything either in as cloud-based SaaS or even a well-established vendor is a nightmare of risk assessments and red tape.
I think this type of tool is very valuable for a lot of these firms but the ones that need it will be the ones that have teams of people trying to block you. One answer might be SaaS-based but hosted on-site to avoid the security issues people have around having their IP in the cloud.
Best of luck!
We have had a couple of larger firms suggest this as an option they would prefer, but moving quickly in this direction is the biggest problem though. Thanks!
Have you got anyone in finance actually using this at the moment? I imagine there's a pretty high bar to entry in terms of security/stability for them to even give a product like this a test drive.
Finance here. Just gave it a try, it won’t fly. There are 2 types of finance: those that take a spreadsheet as it is and just “use” it; those that read formulas. Neither of the 2 is going to get any value from this. On the other hand, finance has a lot of drill downs going on (think going from global to region to country to entity etc.) and I like the way he is displaying data trees going out of Excel for it.
We think a lot of the value in Slate will lie in the ability to alter the level of detail you can see in the model. One of the really exciting ideas we have is to allow you to group together nodes in a tree to form a "parent node" representing some higher level functionality (e.g. by geographical region as you suggest, maybe EMEA forecast for example) . Then there will be the option to take the general overview of the spreadsheet, or dive down into the details.
Hi, then if you are building this you should put a real database in the middle and work on hierarchies, structures, mappings, comparisons and drill downs. You should allow multiple users to pull data from said db into their Excel. You should do this better than the existing (crappy) solutions like HFM or BPC. If you do that, I see quite a lot of value for finance-controlling.
- Getting the license key was a bit slow, I went through the process of getting a reminder before the original turned up (I guess I'm impatient - but it's annoying when I had to hunt around my Junk Mail/Deleted Items to see where the license had gone). Why bother with a license for a demo?
- I find the appearance is too different to Excel so is a bit jarring
- The animations are cute, but not sure that they actually aid usability
- Although my initial reaction wasn't that positive it did actually grow on me but I think it needs an obvious immediate impact to demonstrate a clear benefit over the standard Trace Precedents/Dependents
- I can actually think of one scenario where this would have been pretty useful. I was designing a industrial modelling application that had been "prototyped" as an extremely complex Excel application - I had to spend a lot of time showing people that the calculations in the spreadsheet weren't what people thought they were - so having some way of automatically generated documentation of parts of a spreadsheet would have been useful.
Overall - not bad, but I think you really need to nail the exact pain points this would be useful for. Maybe also drop the Tony Stark/Avengers example - it's cute but might not appeal to your target market.