Zapier Usage Research

In July of 2013 I did some business related research regarding the type of usage made in Zapier. The research involved scraping the list of zaps that’s publicly available on the Zapier website, dumping them into a database, and fetching a few queries with different intersections on the dataset. The business I was planning at the time is now acient history, so I thought I’d share my results with the world, in case anyone might find it useful.
*This article is part of a business plan I decided to edit into a series of blog posts. You can find the rest of the content here

Patents Are Going To Change

Within about two months time from the publication of this post, the patent for the portable computer is going to expire. Considering the proliferation of laptops, and the amount of patents involved with the manufacture of one, it’s hard to predict whether this change is going to affect the industry to any extent. However, this does illuminate clearly the point in time we’re in within these early stages of the age of information. It’s well known that the patentability of software was established within the U.S during the early 90’s, and that they’ve been growing in numbers ever since. This is important for us, because these days the first few batches of computer related patents are being released for free use. It is the unique character of this industry that makes this advancement un-newsworthy, because of a perceived pace of progress under which nothing of 20 years of age could be of any worth. Anyone who’s been around that long, though, knows this to be untrue. Innovation works much the same as a pendulum, moving from one extreme to the other. Such is the way Anil Dash describes the move from mainframes to the personal computer as a revolution of sorts, only to be followed by a swing of the pendulum in the opposite direction with the proliferation of the cloud. The same could be claimed with regard to IDEs, visual programming languages, imperative programming paradigms, and many other aspects of technology that we consider as innovative. Today’s world of technology is without a doubt on the move forward, but much of the conceptual foundations of today’s inventions were already laid out in legal speak in a previous turn of the wheel. This has the dual effect of allowing us to move forward in fronts that were blocked from extensive (as opposed to intensive) progress up to now, and also, limiting the possibility of a future appropriation of these innovations.

The Opportunity

When speaking of an extensive progress that weren’t possible up to now, one can imagine a world of smaller and smaller manufacturers, much like the way every town had it’s own bread makers and shoe makers at some point in time. This might seem counter-intuitive with regard to a common rationale of centralized manufacturing that can cuts costs, but this isn’t really the case in most industries. Nike has centralized manufacturing, but their products cost a fortune, because they have vast expenses for the purpose of brand building, and because there’s enough of a demand to make it worthwhile. The telecommunications equipment industry is another kind of example. A few manufacturers like Cisco have vast expenses in R&D and they aim for the higher end of the market. However, most of the consumers are service providers in developing nations, where infrastructure has yet to be deployed. In these markets there seems to be a clear preference for smaller manufacturers that provide cheaper goods. Another result of the same process could be better kinds of free software alternatives to existing products, and in particular a preference for maker alternatives or pretail variations that could replace the mass-manufactured goods of today’s world. The rise of 3D printing and maker culture saw an instantaneous reaction from IP aficionados that claimed for the need to implement IP protection mechanisms within printers, even for non-commercial use. But the expiration of patents for much of the technologies that are simple enough to be of relevance to makers would change the balance of power in these regards.

The Future Of Patents

Today’s innovative sphere is advanced in two parallel trajectories. On the one hand are entrepreneurs, hackers and makers, and on the other, a narrow oligopoly of tech giants. In a recent talk with Israel Twito, CEO of New-tone Patent Search ltd., it was noted that the same balance of power between big manufacturers that prevents smaller players from being able to compete, has the double-sided effect of forcing these same companies to continue the arms race forward, with intensive R&D, as well as through the constant acquisition of smaller players. When passing criticism over an exits-focused startup culture, one has to keep this state of affairs in mind. Some industries are blocked from entry to anyone that isn’t deep-pocketed enough to join into an arms race that’s already under way. On the other side of the market are big players that need sources of producing new IP, with little or no regard to the cost – both the financial one, and the one that has to do with the damage to innovation and to clients of acquired ventures. This saddening state has an upside as well, one that will only begin to show within the next few years. The tech giants are arming themselves with patents for their own ends, but are also making the patent system obsolete by way of doing so. With the realization that a patent’s strength lies in the abstractedness and generalization of its phrasing, and with the constant swings of the technological pendulum, we’ll see more and more areas where the field for innovation is ironically already covered by existing, expired patents. With that process underway, the tech giants of today will be forced to create products of such sophistication that their target audiences will narrow, much like Cisco’s does today.

An End Note

The grinding of the patent system by economically motivated and heavily funded players requires new and better tools to help us understand what opportunities lie before us. A survey of U.S patents that are about to expire within the next year shows interesting things like cloaking systems, software protection machanisms, electric car chargers, pen-based computer inputs, home networking, and a lot of the technologies around cellular communications. Crossing that information with patent lawsuit databases shows that not a lot of today’s expiring patents are inhibitors to innovation (or competition), but that’s going to change at some point in time, and we best be prepared.

My Cofounder Video Marketing Endeavors

I prepared a short wanted ad for my next venture with the kind help of my beloved partner. It got me to the first page of Youtube’s search results for the keyword “find a cofounder” with just friends and family marketing. It didn’t get me a cofounder. This indicates some relatively obvious conclusions: (1) There’s ad arbitrage on Youtube, and (2) this doesn’t matter if your target audience isn’t looking for you over there. Either way, here it is:

PinMyScreen – Bookmarks Using Pinterest

PinMyScreen is a Social bookmarking extension for Google Chrome. The extension adds a Pinterest pin button to the browser’s toolbar, and allows the user to pin website screenshots onto his Pinterest boards. This works the same way as with gimme bar only it doesn’t impose a new system on clients. Instead, PinMyScreen allows you to use your existing Pinterest acocunts to save your favorite websites.

Known Issues

PinMyScreen uses Chrome’s built-in screenshot functionality (the same one that’s used for the ‘most visited’ page you see when opening a new tab. This, for security reasons, prevents us from screenshoting any sensitive data that might exist within secured pages or Chrome web applications. As a result, not every website can be shot, and you’ll sometimes find that the extension sends an empty image onto the Pinterest ‘pin this’ form. The process of screenshoting a webpage requires a roundtrip request and response with our server. You should note that we do not keep any images or any other sort of information on those servers or anywhere else. As opposed to many other Chrome extensions, I’ve chosen not to use anonymous analytics as well, so no data on your actions is logged at any stage. However, communication with our servers sometimes takes a bit longer than expected. You should note, that if you clicked the ‘pin this’ button on your browser’s toolbar, the request was sent and the pinterest form will appear in a few seconds time. This however doesn’t always happen fast enough, and there isn’t an indication that the request is in a processing state. If I’ll hear a lot of complaints from people about speed, I’ll transfer the servers onto Amazon’s cloud services, so if you think that there’s place for improvement, just let me know.

Bug Reporting and Support

I’m always available to hearing feedback over the extension, be it a positive or negative one, so don’t hesitate. If there’s anything you need help with, just send me an email to me at shayacrich dot com.   Visit our page on the Chrome WebStore to add the extension to your browsers.

Developer & Vendor Tools for Data Integration

Imagine a scenario where you launch a new project management tool, and it’s really the best one out there, but no one wants to use it. This could happen for several reasons, one of which would be the lack of integrations with other types of software your potential users are using. They might need integration with Google Calendar, or with a variety of CRMs, or ticketing systems, or bug tracking systems, or version control services, and so on and so forth. Eventually, you’re forced to start providing built-in integrations with any and every other tool out there. With each new integration offered, the potential user base of your app grows bigger, which is good, but this scenario has it’s disadvantages:

  • There are infinite potential integrations, so the work is unending.
  • This diverts the attention of your development staff from what they do well (project management, in the case described above) to other types of work.
  • Bugs are usually hard to control because you’re working against systems that you don’t host, and whose code you can’t see.

The need for integration tools aimed at software vendors derives from the scenario described just now, which I like to call the integration pitfall. Application providers are faced with pressure, having to offer built-in integrations with a growing subset of related products, so that their own product would be useful to new target audiences. Some tools were developed with the aim of assisting these software providers, instead of targeting their end users directly. is one such tool. It lets you connect with a variety of third party services and map fields from these data sources to your own application with a simple drag and drop interface. In sum, is meant for non-technical users that want to extend the application that they offer with built-in integrations. A very different approach is that of Nectil, a French service provider that developed their own software for web application integrations, and that offer small businesses tailored solutions on top of that framework.

Most other tools aim at the developer crowd, trying to help them consume APIs with greater ease and speed. Temboo provides SDKs for server side API consumption. It supports a very extensive amount of APIs and provides access to them through a unified interface. The service is SaaS based, though cheap, and meant for low levels of consumption (anything really big would require custom pricing). Webshell is an API that lets you integrate different APIs from third party services and create new functionality. It’s a JS library, which handles authentication for you and lets you construct your custom APIs using an editor. DERI Pipes is a software that lets you create automated processes that transform and mash together web content (much like Yahoo! Pipes, but more technically complex).

Cumula is a PHP application framework that takes the idea even further. Based on the assumption that modern web applications no longer sit on top of a single on-premise database, but instead connect with a multiplicity of services, they’ve contrived a framework that lets developers build applications with a unified interface for local and remote data sources. It’s modular, with the use of components for describing parts of the application. Each component has its own routing and templating. A collection of DataStore and DataService packages are available as dependencies, included in the same Github repo. As of now, Github’s popularity metrics indicate very low acceptance rates, and the same goes for the associated Google Group. Nevertheless, this is a relatively new project, and they’ve already received some media coverage.

Other platforms aim to help developers find and consume APIs, without altering the way they work or requiring the utilization of a specific framework. APIHub (renamed to the anypoint portal after being acquired by Mulesoft) and the ProgrammableWeb projects are API explorers, within which one can quickly search and learn about new APIs. Mashape does the same, although focused more on developer APIs and less on connecting with third party applications. Mashape acts as a middleman between providers and consumers, relieving API consumers from having to maintain separate authentication processes with a multiplicity of remote services, and thus enabling the proliferation of usage of such services.

Edit: Since the time of writing this, a new contestant has emerged, called CloudElements, that provides a uniform API for accessing multiple cloud based applications.

*This article is part of a business plan I decided to edit into a series of blog posts. You can find the rest of the content here

Guidelines for hiring someone to build your startup

1. You should have a very clear understanding of what you need built. Usually, mocking up the wireframes and writing a short spec that specifies the processes users go through in the system is a crucial first step. If you can’t do this yourself, the developer or team you work with will have to be experienced with doing this for entrepreneurs, but I definitely do recommend you try to do this yourself. 2. The developer or firm you work with should issue a detailed estimation and timeline, so that you’ll have a clear view of what’s going to happen. This will allow you to change course during development if you feel the work isn’t progressing in the right pace or if you need to change the requirements once you start to see your vision realized in code. 3. You should make sure that you’ll have the ability to track the work on a weekly basis, or sit with the developer in the same room. This means you know at any given point in time how many hours were already spent, and how many features are done. This will allow you to control the level of fine tuning of the implementation of the design and UX, as well as skip features when you fear the system might not be ready before your money runs out. 4. If you’re burning your own money, you should have a very very clear vision of how you intend to progress once the MVP is ready. Remember that an MVP is not a finished product, and there’s a limit to how much you can monetize a product at that level of maturity.

The Super Distribution of Data

A recent article on ALA pronounced the proliferation of content aggregators such as RSS readers or mobile apps like Flipboard, as ‘the future super-distribution of content’. Readers want control over formatting, frequency of consumption and methods of sharing, and they’re turning away from the content providers and using specialized tools that fit the uniform experience that they need.

For most of us, this is already a given truth, and looks like a rather obvious progression from previous technological advancements. However, at the same time, we also take for granted that the web applications we use for our day-to-day trade are contained silos, with their own special UIs and terminologies. Each of these contained silos holds a segment of our data, and forces us to sign in so we can view it, and to adapt to their idea of how to present our data to us.

Gradually, businesses are adopting data aggregators that simplify specific segments of their day to day work. For instance, a business that relies on PPC-based marketing, might start using a PPC management tool, from which to manage their entire ad spend from a single user interface. Now, they no longer have to sign in to multiple services, and cope with the varied complexities of each system.

With the adoption of such tools, comes a new expectation – that other aspects of their business be managed with the same simplicity. Each aspect of a business’s operation deserves a specific tool that would abstract away the complexities of the underlying systems. But then arises a new multiplicity of tools, which suffer from the same sorts of problems – multiple sign-ins, multiple UIs, and data that’s contained within specific apps, not usable anywhere else.

I propose a bigger revolution – one of a super-distribution of data at large. The principle logic behind this change is the recognition that building a truly one-stop-shop for any and every need of a business is unachievable. Instead, one should focus on enabling the distributed model, in which multiple complementing services are used in parallel. With this recognition in mind, I wouldn’t suggest to anyone to try to replace the PPC management tool mentioned above, but to complement it by enabling it to communicate with other tools meant for other specific aspects of a business’s operation.

* This article is part of a business plan I decided to edit into a series of blog posts. You can find the rest of the content here

Excel Spaghetti

The need to combine different data sources that exist only within cloud applications forces many small business owners to turn to working with spreadsheets. These, have several advantages that complement the standard ‘menus and tables’ UI common in most web applications:

  • The ability to view all the information from one screen.
  • The ability to change values on a whole column in bulk, instead of having to go through each item separately.
  • A UI that non-technical people are familiar with and understand.
  • The ability to merge data from different sources onto a single spreadsheet.

In spite of the apparent comforts of managing a business with spreadsheets, this has the disadvantage of being a completely manual process, which has to be repeated at intervals, including a preparatory stage of merging separate spreadsheets into one. The outcome of this work is what I like to call ‘Excel Spaghetti’, as more and more files are accumulated on the desktops of the employees responsible for the manual labor, slowly contributing to an unmanageable work process. The apparent disadvantages of such a manual process would be:

  • The lack of version control mechanisms, and a lack of a single source of truth.
  • Work is saved locally and transferred by means such as email, instead of being concentrated in a single location.
  • Data wrangling, as well as re-formatting the different sources of data, is needed each time an update is required, usually over a dataset that’s mostly the same as before.
  • Data is never really up-to-date by the time it’s downloaded and re-formatted for presentation or analysis.

*This article is part of a business plan I decided to edit into a series of blog posts. You can find the rest of the content here

Overview of Process Automation & BPM Solutions

Process automation tools, meant for either small businesses and consumers, or for Enterprises, allow non-programmers to combine services from different vendors into a new piece of functionality. Enterprise and consumer tools differ from one another in the level of complexity and technical know-how that they require. Most tools meant for consumers can be configured with no more than a few simple steps. They allow people that don’t write code to define automated behaviors between cloud-based services, such as sending yourself an email when this or that happens, adding a new contact to your mailing list or CRM when a form is submitted somewhere, or sending new content directly to your Dropbox folder. These can be regarded as productivity micro-tools, and are very limited in the scope of possibilities that they offer.

Zapier (YCombinator alumni) is a prominent tool, meant for very small businesses and consumers, which offers an extensive list of third party services to integrate with. IFTTT (a Betaworks profile company) is another prominent contestant, meant mostly for consumers and free of charge. Cloudwork (sister company of GetApp) covers the same space as well, and are working on a set of SDKs, and a visual editor, to allow anyone to create new integration types. Yahoo! Pipes is a tool to aggregate, manipulate and mash up feeds from third party services, that offers lots of useful functions for processing the data and is all assembled visually. Other alternatives in this sphere include RunMyProcess, Foxweave & WeWiredWeb.

BPM software for Enterprise customers are the other end of the spectrum. They normally provide form designers, visual process editors, rule engines and customizable dashboards for monitoring workflows. These solutions are relatively complex to operate, and usually require a learning period to adapt to. Their pricing levels are clearly meant for the Enterprise world. Competitors in this sphere include Appian, Skelta, Rollbase, Iceberg and ProcessMaker, to name a few.

*This article is part of a business plan I decided to edit into a series of blog posts. You can find the rest of the content here

Overview of Dashboard Solutions for Data Integration

The tools surveyed here offer an integration of data into a single platform, but do so only in one direction. By this, I mean that these solutions don’t offer the possibility of sending updates back to the applications from which the data originally came from. One such tool is Sumall, which offers a marketing-focused set of metrics based on third party data, consumed via APIs. A competitor in the same vector is Geckoboard, offering an API itself for end users to add their own services on top of those provided out of the box.

A slightly more configurable tool is a Chartio, which connects with what are mostly DBaaS providers, to allow one to create his own customizable business dashboards. Chartio’s platform allows non-technical end users to author their own graphs by dragging and dropping fields from the third party data sources they’re connected with. Combined with that, is a rules engine, based on a set of dropdown menus, for setting up the logic behind the presented graphs. A more technical option is that presented by Dashku, offering a platform on top of which frontend developers can build widgets that connect with third party data, and that combined, are gathered into a fully self-authored dashboard.

Every BI software provider offers some level of dashboard building capabilities. Notwithstanding, these solutions are usually out of reach for small businesses. Jaspersoft, for instance, offers a community edition of its solution as an open source project. However, that open source version lacks the dashboard capabilities that paying users are offered.

A subset of dashboard providers is that of analytics solutions. These are less likely to integrate data from different sources and are therefore of relatively minor significance for this overview. In spite of that, some tools are still worth noting. The first of which is Chartbeat, which gathers data from the users’ websites, but integrates with social media data as well, to generate a more insightful perspective, all without requiring the end users themselves to do the heavy lifting data work. Another tool takes that approach slightly further – ThinkUp pulls together data from users’ social networks, but doesn’t offer graphs and dashboards at all. Instead of pouring immense amounts of data onto the screen, ThinkUp presents a feed, much like that of most social networks, that highlights actionable insights based on the sum of the data that was gathered.

*This article is part of a business plan I decided to edit into a series of blog posts. You can find the rest of the content here