You are here

Mail Search Upgrade – What, When, How and Why?

Written for: 
Instance Affected: 
Australia/NZ
Change Scheduled: 
Sunday, 10 May, 2015 - 07:00

Over the last 18 months we’ve been quietly improving Aconex search, starting with Mail search.  We’re reaching an important stage of our project and we wanted to share how these changes will affect you.

The beta of the new mail search is available now - give it a try!

SUMMARY

Why - You expect more from search than you did two years ago. We do too. So, we’re improving searching in Aconex to make simple searches simpler and complex searches more accessible.

What - A significant upgrade to Aconex search

How - Listening, watching and responding to feedback from you, our users.

When

Stage 1 (Now): Our beta Mail search is available to all Aconex users now. Check it out and let us know what you think, we love getting your feedback.

Stage 2 (June/July): We’ll encourage more people to start using our beta Mail search.

Stage 3 (August/September): We’ll move everyone across to beta Mail search (don’t worry, you’ll still be able to toggle back to old search if you’d like).

Stage 4 (November): Beta Mail search will become our new Mail search. The old Mail search will be retired.

Stage 5 (ongoing): We’ll continue to keep extending and improving search.  

 

HIGHLIGHTS

You can now filter most columns directly. You can also set your own personal default columns so that your search setup is always the way you want it.

You can use specific filters for different field types eg: Date, Boolean, Text, Lists.

Query Summary indicators ensure you always know what generated your current result set.

Coming soon -

  • Row per Mail - only show one row per mail regardless of how many recipients there are
  • Row per Recipient - optionally expand one or all mails to show all recipient data
  • Restrict the recipients you view to TO and/or CC.

DETAILS

The following section provides a detailed background of the changes as we know some of you want ALL the background.  You may need a coffee!

WHY

“Why change search? It works.”

Search in Aconex does work. But, since the dawn of Aconex (way back in 1999), much has changed in what people expect from a search tool - especially thanks to our good friends at Google.  

The benchmark for what search can do has shifted and we need to meet (and hopefully surpass) that benchmark in order to ensure you can continue to get to the project information you need, faster and more efficiently.

Nowadays search must be simple and powerful at the same time.  It must be fast and effective, easy to understand – essentially you need to be able to find what you need quickly and efficiently.

Aconex in a nutshell exists to help you record information (mails, drawings, documents etc), share it selectively between project participants and access it efficiently when you need it.  Get stuff in, move stuff around, get stuff out.

The “getting stuff out” bit means search and the quality of our search tools pretty much define Aconex for the majority of people using it. It is one of the fundamental and primary interactions people have with the system and their success with it is core to how effective Aconex is as a collaboration tool on their project.

WHAT

When you search within Aconex you are usually searching for either:

A Needle in a Haystack - there’s one specific item that you need right now,

Specific Monitoring/Reporting - you’re looking for a set of items which match a specific set of criteria, but you’re not really interested in the individual items.

The vast majority of actual searches done on the system match the “needle in a haystack” model but the existing search model is much more aligned with the “Specific Reporting/Monitoring” mode.   Now that mode is very important too - even though a much smaller minority of searches use that mode, they affect a much broader number of people and processes. However it was definitely time to rebalance the model a little to better support the haystack hunters.

So.. the model needs to support both of those modes. Boiled down, it needs to:

  • Make simple searches simple
  • Make complex specific searches more accessible

Needle in a Haystack searches tend to follow this pattern.   

Start with a simple generic search with one or two keywords to narrow down the search results to a small enough subset (normally one page) from which you can spot the one item you’re after.  For this type of searching the important bit is making that first filtering as simple and efficient as possible.

We’re focussing a lot of our efforts on getting you that first set of results fast and to make it easier and easier to massage and adjust it till you can narrow down enough to spot your target.

By contrast the Specific Monitoring/Reporting searches require much more specific queries and tend to utilise the specific filters.  Our focus there is on making the majority of the available metadata directly searchable via the column header filters - no more needing to know the magical supersearch query syntax.  You can create various custom column/filter configurations and save them as either your default column configuration or as saved searches for future usage.

HOW

Over the last 18 months, we have been working on the first stages of our new search model.

We chose the mail module as the starting point for this work as the Mail Forms feature was coming out and all those new customisable fields were going to need to be able to be searched somehow.  

In June 2014 we released the first iteration of the beta search model and quietly invited users to give it a try and tell us what they thought.  Based on that feedback and observations from our ongoing usability testing we’ve been iterating, clarifying and enhancing the new search model.

Initial feedback was broadly encouraging but we also learnt that:

  • Our new column header filters were not displaying clearly or quickly enough, many people simply didn’t see them at all and assumed we’d removed the ability to search via anything other than just the supersearch field.
  • Some browsers and projects were seeing significant performance issues when viewing results.
  • The auto reloading of results whenever a filter was changed was frustrating to users who wanted to add more than one filter at a time.
  • Some of the things we were doing to try and get results to you faster were having the opposite effect in some circumstances.

Armed with that feedback we went back to the drawing board and started chipping away at the rougher edges and adding additional features.  At the same time our engineers were completing a comprehensive refactor of the code beneath the surface to support some of the newer features we have planned for Search in the longer term future.

We’ve been steadily and relatively quietly working on the beta search model, watching the feedback and usage stats and adapting and adjusting the model accordingly.  We’re now almost ready to move to stage two of the beta process where we’ll be encouraging all users to try the beta search model and to let us know what remains to be done before we can switch off the old search.

WHEN

In the next few months we will start encouraging all users to move across to the new search.  The old search will stay available during this period.

In the next stage we will then automatically move anybody who hasn’t yet tried the new search across – they will still be able to toggle back to the old search if they need to, but be aware that by this stage, those days are numbered.  It’s important that you let us know what it is that is making you move back so we can adjust it before switching off old search altogether.

WHAT NEXT

Search is a huge and almost never ending bucket of opportunities for enhancement and there are many things on our todo list once we have migrated to the new search.

  • Type ahead controls - for looking up users and organisations
  • Smart Help - easy access to guidance about advanced search techniques
  • Mail Tagging and Flagging - ability to mark mail as “todo” etc.
  • Conversation view - a search mode that allows you to view the mails from a related ‘conversation’ all together
  • “Contains” searches - so you don’t need to be quite so exact in your queries
  • Searching mail attachments (file names and content)
  • Project level default columns - so you can setup your search environment for everybody in your org/project

… and more

BENEFITS

FUNCTIONALLY

  • More familiar inbox-like behaviour
  • Search results display faster
  • Basic searches are easy to trigger
  • Complex searching is more accessible
  • 95% of all mail metadata (including custom mail forms) is now directly searchable
  • Easy search across all that metadata - no longer need to know or look up complex search index prefixes)
  • All list fields (including single-select fields like Mail Type) are now multi-select search filters
  • You can customise your search screen setup by saving your default columns/filters
  • You can filter by Unread
  • Ability to save your default column sets (have your common fields and filters always available)
  • Significantly reduced clutter and noise with more focus on the content and less on the filters
  • Easy to see which filters you've applied
  • Sticky settings that remember how you left them last time (Number of Mails per page, Show

Mail Only, Default Column Sets)

Coming Soon

  • Row per Mail mode - Further reduce screen clutter and scrolling
  • Search for TO or CC recipients explicitly
  • Highlight the last mail you viewed from the list

TECHNICALLY (if you’re that way inclined)

  • New Internal API logic is more scalable and more accessible to other apps that need to interact with Mail data
  • A much more modularised codebase makes it simpler (and safer) to make changes
  • Adopting progressive tech frameworks to support us through the next five years of ACX development (REST, AngularJS, Microservice Architecture)
  • Significantly improved automated test coverage (from 0% coverage prior to 90% in the UI), which makes future development fastermuch less risky
  • Cleaner code that removed unused/outdated functionality
  • Greatly reduced reliance on the session storage, making the searches much lighter resource wise and aiding general app performance and uptime
  • Ready for next steps, taking advantage of the incoming Elastic Search indexer (due 2016)

Any feedback?

Thanks. A ticket has been opened with the Support Central team.