Ask YC: Review our new product
cleanupdata.com
cleanupdata.com
At the same time, considering that the data work may be large, having this on a server is not too bad of an idea.
What I suggest to you is this - spend a day to make a version of this wrapped in a web browser in a .exe file. Put a price on the .exe, and sell it directly to businesses. It's an easy sell, and it's way more convenient to have a "local" application for a tool like this, than a website.
If the product could be implemented as an extension for Excel, similar to Xobni in Outlook, then there is much more potential. There is really no reason for this to be an online app.
I think markessien has a great point. A lot of enterprises don't want their office worker in browsers (however lame that sounds) and may not be comfortable sending you their data.
Having a web app style wrapped in a local framework would allow you to sell to these companies what looks to be a really great product.
That said please don't loose the web version for those of us that realize it's the 21st century already.
Sure this should be part of whatever spreadsheet you are using. I totally agree with that. But it's not. That's the reason someone came up with an external solution.
But why shouldn't it be a web based app? If it works just as well? If you have privacy/legality issues then OK, you need a local app. A local app may also be just as good for many other things. But why is this less suitable to the web then word processing or spreadsheets?
From there on it's just a marketing/distribution issue.
But I don't think this is a solution to that problem for most people who have that problem.
My previous job involved a lot of Excel work (so I became fairly good at it) and it also involved interacting with people (often clients) who used Excel occasionally and really werent very proficient at using it. So for simplicity, lets imagine the world of excel users is made up entirely of those two types, and lets call them A and B respectively.
Types As don't need this service - they can do all the things you showed in Excel quickly (using keyboard shortcuts etc), using formatting codes (even though Excel format codes are clunky. eg. SSN format code is: ###-##-####).
Type Bs do need your service. But will they use it? If they use Excel rarely then is it worth the overhead of knowing how to use a separate service (even a simple one)? If they can't figure out how to use Excel codes then are they the type of person who will be interested in exploring new services? (note, this is about the type of person, not about how simple your service is)
OK, so the world of Excel users is not as simple as the A/B world that I describe. There is no doubt another type C somewhere between the two who might value this - but how many of those are there? My gut says 'not many', but I don't have data. Can you describe to me your archetypal customer?
Making this an Excel plug-in might solve this overhead problem somewhat, but even then, the latter issue still remains.
The type of person who can't program a blog but can & will install a Wordpress if she can find decent instructions because blogger doesn't let her do XYZ.
The type of person who can & will use forum code but can't use html/css.
The type of person that can & will set up a hosted online shop but can't alter the code.
Basically, if the problem exists & the excel solution is not something that you would just come across in excel without knowing, it comes down to what solution you do come across.As an online app, this might have a better chance of being that.
The second thing is this: would you (A) prefer to explain to your clients (B) how to use the excel solution or send them a link to this?
Tip of the hat to LAPIS/Potluck/MassEdit from CSAIL for getting us thinking in this direction.
Your only goal should be to try and sell yourselves to Microsoft as quickly as possible.
While clever, the central idea is trivial to replicate. Try obtaining as much of a legal protection as possible. You have a year from your initial public disclosure to file a patent application. This will run you at about 10-15k, but will greatly increase your chances of a technology acquisition. If in doubt about the patent, consider the fact that once a variation of this gets into the next Excel version or a service pack, your customer base will go to zero.
Definitely drop the SaaS model, this is the worst application of it ever. Virtually all business data is confidential, copy-pasting it around random web servers is not encouraged to put it mildly.
It is a nice idea and it is certainly a neat tech demo. Your best bet in my opinion is to try and sell the actual technology bit.
What market are you going after? Soccer moms managing team phonebooks? Entrepreneurs who have other things to worry about than cleaning up data?
It'd be cool to see see prompts that inspire better uses/structure to your data. E.g. if a company is managing their sales pipeline (as in the dabble db 8 min demo), analyze that data further than excel easily allows.. track how the data changes over time to deliver insights that people wouldn't see by just producing graphs. What common trends exist among sales that actually close? How long does a sale stay on the table before closing? Does this vary by salesperson/location?
It makes sense for people who know little about spreadsheets [like my mom who would manually add dashes to dozens of rows :) ], but they may also be the type who wouldn't think to seek out a web app to help them clean up the data.
I'm probably a poor sample b/c I spend 5-6 hours/day in excel (finance @ tech company).
Color-logo feature is sweet!
However, for those advanced use cases (when it's not a one-off), there is a need for saving/editing/copying/versioning of transformations. Are you guys thinking about providing such capabilities?
I don't think it's patentable[1]. It's a cute idea, kind of obvious; I'm not sure how well it really generalizes on all the cases that come up in huge datasets. If you continue along this path, you might come up with some patentable detail. I think in the US, you can file up to a year after inventing, so public demos don't destroy novelty (IANAUSAPA).
I think your strength is less the technology than the interface, which is nicely wrapped up and works[2].
You'll get interesting user leads from this demo. Maybe: who is suffering from what problem in what specific context who really needs your help. It's hard to get this market research any other way,
I can't see a really neat or natural way to monetize it - it's just not big enough. But the world's imagination is greater than my imagination. If you stick at it, I think something will come to you. That would be really cool!
[1] a patent isn't for suing people, but to have something to sell to Microsoft or Google etc.
[2] of course, you can improve the GUI further once you know what people need - imperfection is a good thing, because a head start on the competition is worthless when you run out of track.
Suggestion for getting more helpful feedback -- the first post on this thread should have been this:
My name is Avi Bryant and I'm the chief bottle washer at X. We are going to release a product which performs automatic transformations on data, with an easy-to-use interface based on building examples.
For example, let's say you have a data entry team which has painstakingly transcribed 100,000 customer entries over the years. Their phone number field is an absolute mess -- some customers write 4567890, some write 123 456 7890, some write (123) 456-7890, etc. You want this to be consistent. You could either hire a DBA to do some database/regexp wizardry for you, or you could just write a single example of the proper canonical phone number and our software will just work.
For an example, try playing directly with this test data: http://www.example.com/i-go-to-test-data-not-a-video
We would particularly appreciate feedback about ease of use and ...
What's your business model for this? Are you just doing advertising or are you also offering a rush fee or maybe charging extra for files over 5MB or something?
You could keep it as a web service and/or charge to license it to businesses and make residual income off support/updates/installation on their network.
1. the icons that appear under a column when you're editing it should have tooltips so you can see what they mean
2. there should be an undo function
3. it should be able to figure what you mean if you retype as upper-case rather than using the buttons. (Perhaps asking the user "do you want to put it in uppercase"). This is not exactly hard to code, e.g.
if (toUpper(originalField) == editedField) assumeUserWantsUppercase();1) I could easily paste like that (most of my data is large and too big to be pasted)
2) Since the data I have has been collected under a Privacy Policy, I can't upload it to other sites.
But for small jobs of flipping things around, this is a cool app.
I do this kind of manipulation quite a lot, but use my lovely emacs to do it. snip snip - easy peasy. But I know it's not reasonable to expect regular excel users to use emacs.
Feel free to include the code in your tool if you want to offer MySQL as another download option when users get their results emailed to them.
(I copied your pasting instructions. I hope you don't mind.)
But if it was an Excel add-on of some kind I bet you could sell it. Maybe to Microsoft. Or, sell the underlying technology to Google Docs. It's got a wow factor that could inspire upgrades or switching.
Anyway I doubt you hadn't thought of that already. Cool stuff.
The amount of data will nearly always be tiny so you ought to just process and display it interactively unless the dataset is huge. And speaking of interactivity, it really needs an 'undo'.
I'd like to read a blog post about the details of how you did it.
Just curious though, why does it have to be a web app? Why not just a plugin to my favorite spreadsheet application, or just a small standalone app that I can use on my PC..
You guys looking for an intern? lol