I've figured out how to implement the automatic detection of document type by scan - efficiently and quickly. Without spending $150K and 4-6 months at least. It will be implemented soon.
1. Be curious and ask questions about what users do 2. Run experiments all the time 3. Make incremental changes 4. Test and iterate 5. Be ready to fail, but desperately look to win
Last week, we began an active #CustDev for @GetXDocs.
The aim is to gather experience and opinions from professional translators from 10 target countries: Israel, USA, Canada, Spain, Germany, France, Italy, Portugal, Poland, and UAE.
Short notes on @GetXDocs results of the last week.
โ Implemented some more basic marketing mechanics. ๐Fixed one of the core algorithms to reduce the number of server requests by 10 times. โฉReduced the time of translation delivery from 30 to 10 min.
Apparently, there's no API for #transliteration with acceptable quality๐คฆ๐ปโโ๏ธ Again, transliteration is happened to be more complicated than #GoogleTranslate
The #startup I'm working on is a #translation service. There are plenty of different translation services worldwide. But I think we're the first who use #GoogleMaps ๐ for the translations :)
#buildinpublic Today, I planned to give access to my #startup to the first ๐ clients. But couldn't due to the one significant bug ๐ชฒ. Some parts of the system work, but a specific part which is necessary for those clients - doesn't. โ๏ธโ๏ธ
#buildinpublic To maintain the pace while developers are dealing with the bug, I decided to invite 10-30 other clients for whom this feature is not important. Tomorrow, I run ads which I was planning to do in October. So, I'll test another hypothesis quickly โ๏ธโ๏ธ
Today discussed the status of the development with the programmers. There's one feature, important for #MVP, that will be released by Monday - and I'll invite the first beta-testers!