Is it possible to build a working business application in an
hour? Low-code app development platforms like Appian Quick Apps and others are
designed to do just that, but how well does the app work on desktop and mobile?
How advanced are the features? Is there actually zero coding involved? Are
these app builders geared more toward everyday business users or developers?
What kinds of tasks are these apps suited for, be it basic collaboration and
project management or more complicated use cases?
How do you build an app without any coding, anyway? This new
mode of app creation comes with a lot of questions, and PCMag put the idea to
the test.
We took four low-code tools (Appian,
Microsoft
PowerApps, Salesforce
Lightning, and Zoho Creator),
put them in a room with four developers from our Ziff Davis Tech team for one
hour, and watched them try to build a basic app and customize it—without
writing a single line of code. Oh, and we filmed the whole thing, too.
A Little Background on "Low-Code"
The term "low-code app development" didn't exist until a few years ago but the concept isn't a new one. Part of the value for businesses is in citizen development. There's long been a notion in enterprises and small to midsize businesses (SMBs) of the "power user" or "citizen developer": the business users taking it upon themselves to create their own apps, often dabbling in Visual Basic for Applications (VBA) programming in Excel.
The term "low-code app development" didn't exist until a few years ago but the concept isn't a new one. Part of the value for businesses is in citizen development. There's long been a notion in enterprises and small to midsize businesses (SMBs) of the "power user" or "citizen developer": the business users taking it upon themselves to create their own apps, often dabbling in Visual Basic for Applications (VBA) programming in Excel.
The other side of the equation is traditional developers and
IT, for which these low-code platforms are designed to accelerate software delivery by quickly building apps for specific business use
cases. The term "low-code" itself comes from tech research and
analysis firm Forrester Research. Analysts Clay Richardson and John Rymer coined the term in
Forrester's 2014 report,
"New Development Platforms Emerge For Customer-Facing Applications,"
and followed that up recently with two market reports, "The Forrester Wave: Low-Code Development Platforms, Q2 2016," and "Vendor Landscape: The Fractured, Fertile Terrain Of Low-Code Application
Platforms."
Forrester approximates that vendors generated a minimum of
$1.7 billion in revenue during 2015, and said in the report that many are
growing their revenues in excess of 50 percent a year. The Wave report breaks
down 42 vendors in the space, organized into five low-code categories:
general-purpose, process, database, request-handling, and mobile-first, though
Rymer told PCMag that the mobile-first category will likely disappear, becoming
a given in all low-code platforms as the industry consolidates under
general-purpose.
"A lot of people still think of these products as just
tools; the latest turn of the screw in what we used to do in [IBM] Lotus Notes
or [Microsoft] Access. These are not tools, they're platforms. If you're going
to invest in them, you want to be able to create an application in as many
scenarios as possible," said Rymer. "Mobile isn't a product, it's a
feature. The general purpose category has coverage across mobile and web UI
[user interface], tooling to help you manage projects, application lifecycle
management [ALM], portfolio management, administrative tasks. These are very
broad feature sets."
The Forrester report also addresses three key myths about
low-code platforms. Rymer talked a bit about each one:
Myth #1: Low-code platforms are only for citizen developers,
not pro developers.
"The foundation of this research was a collection of
reference customers using these products. When we stepped back and thought
about who these people were, they were all professional developers. We've since
encountered a broader population of citizen devs, but I wouldn't put these
platforms in front of a citizen dev without some training and tooling,"
said Rymer. "I can't think of a product that has very been successful in
satisfying the needs of pro developers and end users with the same feature
set."
Myth #2: Low-code platforms eliminate the need for any
programming (low-code vs. "no-code").
"When we did the Wave on low-code platforms, one of the
questions we asked is when you used these products, what were the functions
that required you to code? The answer was 1: integration and 2: user
interface," said Rymer. "Integration is hard. It's unpredictable.
It's messy. The other areas where people do custom coding is mobile UI. If you
want the app to go beyond the web application and do a custom layout, access
servers on the device, or want pixel-perfect displays, low-code platforms
usually don't provide templates for that. On mobile there's still a lot of work
to do. We've got to get to native."
Myth #3: Low-code platforms mean small scale.
"For me, the really interesting thing about the
landscape is when do we reach the point where acquisitions start to happen?
We're not there yet, because the revenue isn't there except for Salesforce and ServiceNow ,
who are already big vendors, but Appian is over $100 million in revenue...would
somebody buy them? Would somebody buy OutSystems or Mendix ?
We've also got to keep an eye on Microsoft's entrace into the field, because
they could become a huge player overnight. Within the next 2-3 years, I think
we'll get to the point where the Oracles of the world decide to acquire the
leaders in revenue, rather than build. A $50 million acquisition of a low-code
vendor is peanuts to Oracle," said Rymer.
PCMag's Testing Methodology
For the purposes of our test, each developer used their respective low-code platforms to create a basic scheduling app. The goal at the end of the hour was to build an app that could add a new event (name, date/time, duration), invite users to the event, a save button to create the event, and the ability to view a list of events in calendar view or chronological list. If the devs accomplished all that, they could experiment with more UI customization or bonus features like notifications.
For the purposes of our test, each developer used their respective low-code platforms to create a basic scheduling app. The goal at the end of the hour was to build an app that could add a new event (name, date/time, duration), invite users to the event, a save button to create the event, and the ability to view a list of events in calendar view or chronological list. If the devs accomplished all that, they could experiment with more UI customization or bonus features like notifications.
From a feature perspective, Rymer said low-code platforms
replace coding largely with declarative development: drag-and-drop visual
interfaces, object mapping and process modeling, form builders, WYSIWYG
editors, etc. In Appian, Microsoft PowerApps, Salesforce Lightning, and Zoho
Creator, the developers used these types of features to build their apps. The
other big difference between the platforms is that Appian is a low-code and
business process management (BPM)-specific vendor, whereas Microsoft,
Salesforce, and Zoho offer their tools as part of larger ecosystems, with Zoho
also offering products like Zoho CRM , Zoho Projects
,
and Zoho Books .
Why did we use real developers for this test instead of
run-of-the-mill business users? A few reasons. As Rymer explained, developers
and IT are using these platforms in enterprises with far more regularity than
citizen developers. We wanted to test whether, for the purposes of a quick
ticket or light feature request sent to the IT department, building an app
quickly using a low-code platform would be easier than a traditional
development process.
Developers are also far more knowledgeable about what it
takes to create a working app in the first place. In the videos below, the devs
were able to clearly articulate what the platforms could and couldn't do, what
their limitations were, and whether a tool like this is actually helpful
everyday business settings.
Will we run a test like this with citizen developers as
well, or pull unsuspecting business users walking down the hallway into the
labs, sit them down, and make them create a low-code app? Possibly. We'd love
to hear your feedback. Leave a comment on the story, drop us a Facebook comment
or tweet, and let us know whether you want to learn more about low-code through
this kind of content.
How the Tools Stack Up
Once the clock started and the app creation began, the developers found strengths and weaknesses with each low-code platform, but by the end of the hour each one had built an app. How well did the platforms work, and what did the finished products look like? Watch the videos below and find out.
Once the clock started and the app creation began, the developers found strengths and weaknesses with each low-code platform, but by the end of the hour each one had built an app. How well did the platforms work, and what did the finished products look like? Watch the videos below and find out.
The Bottom Line
All four platforms created a working (or at least semi-working) scheduling app, but the low-code tools that performed the best overall were Appian and Zoho Creator. For Appian, the combination of Appian Quick Apps and the full-fledged Appian Designer make for a potent duo in creating an app quickly and then layering customization and added features on top. Quick Apps is primarily form-based, and the full Designer lets you map out all the data and objects in the app with the drag-and-drop Appian Process Modeler. It's the most mature, easiest to use low-code platform we tested for creating BPM apps.
All four platforms created a working (or at least semi-working) scheduling app, but the low-code tools that performed the best overall were Appian and Zoho Creator. For Appian, the combination of Appian Quick Apps and the full-fledged Appian Designer make for a potent duo in creating an app quickly and then layering customization and added features on top. Quick Apps is primarily form-based, and the full Designer lets you map out all the data and objects in the app with the drag-and-drop Appian Process Modeler. It's the most mature, easiest to use low-code platform we tested for creating BPM apps.
Zoho Creator performed admirably as well. Our developer was
able to create the basic scheduling app pretty easily within about 10 minutes,
and then spent the rest of the hour attempting to customize the app. Echoing
Rymer's main limitations with low-code platforms, the development did run into
roadblocks when it came to customizing the UI and optimizing the layout for
mobile. Zoho Creator is the "highest-code" platform of the bunch in
terms of additional scripting on top of the drag-and-drop development and form
building, but packs enough great low-code functionality out-of-the-box to build
a solid app in an hour.
Salesforce Lightning also performed very well, building a
good looking app with a relatively pain-free development experience. The
biggest criticism our developer had with Salesforce Lightning is the
proprietary nature of its ecosystem around the Salesforce App Cloud and its customer
relationship management (CRM) software. In his experience, the
nature of developing an app in Lightning relies so heavily on knowledge of
Salesforce's APEX programming language when it came to customization, he
recommended the product is best suited to developers and users with deep
Salesforce expertise. Salesforce does cover its basis on this point, though,
offering an in-depth training website called Salesforce
Trailhead with a wide array of courses to get you up to speed.
Then we come to Microsoft PowerApps, Redmond's free tool and
the newest low-code platform on the block (it still currently sports a beta
tag), and the one with the most maturing to do. The PowerApps UI is the
sleekest of the tools we tested, and the platform comes with step-by-step
instructions to create and app and a variety of mobile layouts.
Our developer found the basic form UI straightforward and
built the app template in 5-10 minutes, but discovered the program only works
with a Windows 8.1 or Windows 10
machine, and had trouble pulling in data, even when using Microsoft
OneDrive. Forrester's Rymer wasn't suprised that PowerApps came up
short in initial testing, but sees big things on the horizon once Microsoft's
tool matures, especially if they combine low-code with business
intelligence (BI).
"Is PowerApps for end users like SharePoint Designer
was supposed to be, or is it a developer tool? This is one of the big questions
they've got to answer," said Rymer. "Also, I think Microsoft is
eventually going to put PowerApps and Microsoft
Power BI together, with the idea that they would be a twin pair of
tools people would use in tandem. We'll pay close attention to that when
PowerApps goes GA [generally available], because that could be
significant."
Source: pcmag.com
0 comments:
Post a Comment