Golden Rules of Technology in Digital Marketing

Just because it can be done in an app, doesn’t mean it can be done in a website
of course, apps have an additional barrier of persuading the user to download them
Apps have access to almost all features and resources of the device including camera, sensors, Wifi, Bluetooth, full internal storage capacity, and user’s data such as contacts or photos.
Websites are restricted both by the limited functions offered by HTML5 and JavaScript, and the limited speed with which programs written in JavaScript can be run, meaning that anything involving intensive mathematics or large amounts of RAM such as realtime analysis of video is simply not possible.
HTML5 offers some limited access to device features, such as location and user data, but this is always limited. For example, access to photo library is by one image at a time from user selection. Access to location requires user permission and can only be retrieved when the website is in the foreground of the UI, not when other apps are being used.
Not all Photoshop effects can be reproduced in CSS
Photoshop is designed to allow unlimited graphic design possibilities. As much as possible, we seek to implement designs using CSS rather than rasterised bitmaps, this helps with scaling, device independence, SEO, interactivity and more. Therefore when using Photoshop to design for websites it makes sense to limit what you use to a subset of effects achievable in CSS. Even the drop shadows and font smoothing algorithms are different. Similarly, be conscious of fonts, for example one of the world’s most popular fonts for graphic design, Gotham, is not a ‘websafe font’, i.e. it is not included as standard on most people’s computers. Therefore the user may not even see your design as intended, or we would need to include the font in our website code, incurring both a license fee, and dramatically increasing page weight.
There are plenty of things that computers find hard
and thus require millions of pounds of research
Computer vision – knowing what an image is of or about. Optical Character Recognition is about the only viable technology in this space, and that only works by comparing to existing known fonts.
Voice recognition – although some operating systems now do it for you so you don’t have to roll your own (notably, Android)
Learning and self improvement through analysis of historical performance – this can be done but is hugely complex
Augmented Reality does not mean scanning QR codes
Augmented Reality means adding augmentation to your view of reality – such as a heads up display that tells you which direction you’re facing, or a PlayStation EyePet who responds to being kicked or petted even though he doesn’t exist in the real world. Magic Leap and Google Glass are examples of augmented reality. As is Wikitude or LayAR. Scanning reality to figure out what’s in reality, such as scanning a QR code or other symbol, does not make it augmented reality, unless you then use the information gleaned from the scanning to add a layer over reality that complements what you just scanned. Mobile Medic is not really augmented reality, as the view on the phone does not include any element of reality.
All software development takes a long time
Of the following activities, software development ones have the longest lead time. The absolute minimum viable time for a software project of any substance to be done to sufficient quality is 6 weeks. Most campaign work can be expected to take 2 to 3 months. Most enterprise grade work will take 6 to 18 months.
In order of increasing complexity, cost and duration.
Print ads
Web banner ads
Email newsletters
Campaign websites
TV commercials
Enterprise website builds