Tip:
Highlight text to annotate it
X
MAILE OHYE: Hi.
I'm Maile Ohye.
I work at Google as a developer programs tech lead.
This video is one of several in our checklist for mobile site
improvement.
I decided to make these videos in Q4 of 2013,
because I found the mobile web's inefficiencies fairly
disheartening.
To improve matters, in this video,
we'll go beyond the basics of stop frustrating your customers
and proactively work to optimize the top mobile tasks
on your site.
And interesting idea I learned about mobile users is
that just by holding a mobile device,
they're expectation of time can differ
from their desktop attitude.
While site owners were well aware of mobile network latency
and that at its fastest, mobile can't outperform desktop.
"Strangeloop" cites that 85% of mobile users
expect sites to load at least as fast or faster than sites
on desktop.
And when it comes to shopping, mobile users
seem almost eager to close the deal.
55% of consumers using mobile to research
want to purchase within the hour.
83% want to purchase within a day.
And that's exactly why we should discuss
optimizing the top mobile tasks on your site.
Helpful prerequisites for this video include web analytics.
I'll be using Google Analytics, as well as Webmaster Tools
verified ownership of your mobile site.
If you publish mobile content on a separate URL from desktop,
such as on a sub-domain, be sure the m. site is verified.
To optimize mobile task completion,
we first need to gain insight to mobile user intent,
then with this knowledge refine our mobile metrics,
and last improve user workflow.
We'll begin with gaining insight to mobile user intent.
Rather than dive right in, let's take a step back and discuss
mobile intent more generally.
In usability of mobile websites, under what
to include on a mobile site, are tasks
that relate to users avoiding a deadline,
checking rapidly changing information,
viewing private information.
This is probably because mobile phones
are much more personal than tablets or laptops.
The other tasks mentioned for mobile sites
are obtaining business information, like phone number
or store hours, finding directions, and communicating,
such as with email or social networking sites.
Many of these tasks probably resonate
with your own smartphone usage.
Still thinking high level, another way
to understand mobile user intent is through search queries.
Behind each keyword, there's a task
a user was trying to complete.
It may be useful to slice, dice, and categorize your search
queries to provide more lenses for viewing your data.
Informational queries show intent for the searcher
to learn or research something.
Transactional queries reflect the searcher
wanting to perform an activity, such as buying a product
or downloading a file.
Navigational queries are when a searcher
wants to go to a specific page.
Commonly, businesses also include
categories of branded and non-branded keywords,
as well as queries with local or geo intent.
To make things more concrete and to dive deeper into the data,
I'll reference the e-commerce site, the Google Store
at www.googlestore.com in many of my examples.
As for the data, let's begin in Google Analytics
and create a mobile traffic segment.
Start with Create New Segment.
Then select the Mobile Traffic preset.
In Visitors Flow, we can see how mobile users
travel through our site.
For the Google Store, it appears that from our homepage,
about 12% continue to the main shopping page
at /shop.axd/home.
While over 50% drop off.
Although we still need to focus on understanding
the mobile user intent, questions
in the back of my mind at this point
are, why the 50% drop-off.
And is the home page useful?
Additionally, we notice that users like to site search.
To better understand mobile intent,
what are visitors searching for?
To answer this question, Analytics
has a Site Search, Search Terms feature.
Visitors to the Google Store show
intent to see shirts, YouTube merchandise, and also
the Google Store's phone number.
Let's make a note of these top queries.
Given that Site Search shows thousands of attempts
in the last few months, questions we start asking
are, should we consider making on-site search terms more
findable in the navigation?
Can we auto-complete popular query terms
or send searchers directly to the right page rather than
a list of results?
Are the search results even relevant?
It's also fun to look at in-page analysis
to get a sense of mobile user behavior.
When we look at the main shopping page,
we can see the percentage of clicks to each of the links.
Check if the page contains intuitive navigation.
Given common site search queries,
like shirts and YouTube, does the page content and navigation
satisfy the user intent?
Does a person looking for a Google shirt
know to click Wearables?
If YouTube merchandise is extremely
popular with my visitors, should the logo
be included with the main category navigation
in the center or remain off to the right?
And speaking of popular, we can use data to support that term.
Looking at pages within Google Analytics segmented
by mobile traffic, we have a list
of the popular pages for mobile visitors, the average time
on page, bounce rate, and exits.
The five most popular pages by views
seems to confirm a few of the site search keywords.
Google Store's visitors are interested in t-shirts
and YouTube merchandise.
Our understanding of mobile intent
becomes more clear when we investigate web search queries
in Google Webmaster Tools.
Select your verified mobile site,
navigate to search queries, and within top queries,
select a date range of up to three months.
With filters, we'll look at only the searches for mobile phones.
To make sense of the data, as we discussed earlier,
we may want to categorize queries.
The first listed Google Store query variations
are navigational queries.
They require more research to understand intent.
The third query, YouTube, is too broad
to bring us qualified traffic.
So we'll disregard it.
The last five queries are likely transactional with intent
to buy products from the Google store.
Let's further investigate the navigational Google Store
queries to see if we can get an idea of why
mobile searchers want the home page.
This may be especially useful since we already
noticed our home page has over a 50% drop-off.
Still in the Search Queries feature,
filter queries containing Google Store
and keep the search restricted to mobile.
The filtered results show geo-based intent
to find country-specific homepages.
Clicking on the query, Google Store UK,
displays that the page served in Search Results
is still the general Google Store home
page at www.googlestore.com.
This is clicked over 40% of the time,
even though the Google Store has a specific UK homepage
at google-store.com.
A question to ask is, do searchers
intend to find their location's country-specific page?
Or are they desiring to find a page outside their location?
To figure this out, I'll filter the queries
by Google Store UK on mobile with the location limited
to the United Kingdom.
The results show the same number of impressions
as we saw without the country restrict.
We can conclude that searchers seem
to desire their location's country-specific homepage.
In that case, perhaps we can use GPS
to personalize their on-site experience
and show them country-specific content
but keep the tourist experience still available.
You can use this type of approach.
First, categorizing queries and then
investigating more detail with filters, top pages,
average rank, and CTR to further understand mobile user intent.
So far, we've looked mainly at readily available numbers.
But if you have the time, another great way
to understand mobile user intent is to ask visitors themselves,
conduct mobile user studies, or create mobile surveys.
Just be sure the surveys are optional, not a pop-up window
that interrupts the user's workflow.
Judging from the data we've gathered,
let's assume that one of the top tasks for the Google Store
is a user who wants to buy a Google t-shirt.
From our research, we also know some
of the user's common unstated expectations.
For example, if a user intends to buy a Google t-shirt,
they also want the information relevant to them,
likely listing their currency and available shipping.
This became more evident through country-specific homepage
queries.
And the user probably wants to buy a Google t-shirt
with a fast user experience.
This is evident from mobile industry statistics, as well as
the Google Store's high percentage of drop-off.
Visitors are willing to leave before they've even
seen our products.
Now that we've established more knowledge of the user intent,
the second section of the video deals
with refining mobile metrics.
Although, at present, mobile metrics show lower conversion
than desktop, it's clear that mobile research
influences purchasing decisions.
In a Google Nielsen study, 93% of people
who used mobile to research go on to make a purchase.
When refining mobile metrics, it's
helpful to differentiate desktop-initiated purchasing
behaviors with smartphone involved purchasing.
The traditional desktop models include desktop research
to desktop purchase or desktop research to in-store purchase
or desktop research to in-store research to desktop purchase.
Smartphone involvement in purchasing
can look a little different.
There's mobile research to mobile purchase,
mobile research to in-store purchase,
cross-device conversion, like mobile research to desktop
or tablet purchase, or even additional possibilities,
like in-store research, mobile price comparison, and then
in store at a competitor to finally make the purchase.
Given mobile's diverse involvement in the purchase
journey, many find it useful to track micro-conversions.
Within your company, you can report these micro-conversions
alongside the desktop macro-conversions.
They may include when mobile visitors locate a store,
call the store, click the map for directions,
share, save, add to wish list, add to cart.
It's reassuring to note that Adidas with iProspect
found that 20% of mobile users who clicked on store locator
links visited a store.
If possible, also track cross-device workflow with
logged in users.
The main takeaway for this second section
is that because mobile behavior and intent
can differ from desktop, rather than blindly reuse desktop
metrics for mobile, instead look at mobile task completion
and follow mobile behaviors step by step
as a contribution to conversion.
For instance, if a user wants to research and buy a product,
like a Google t-shirt, we could include mobile metrics
of engagement, counting micro-conversions like pages
per visit, add to cart, et cetera, rather than
just orders, as we might do for desktop.
Or if a common mobile task is to obtain a business' phone
number, the metric can be measured
by clicks on the phone number.
Or take the mobile task to find in-store discounts
when shopping.
On mobile, the conversion may be bar code scans, rather than
printing or downloading coupons.
The last section for optimizing mobile tasks
is to use all of the information thus far
to make improvements to user workflow.
To start, actually run through a top mobile task for your site,
using the same types of devices as your users.
The task we selected for the Google store
is to buy a Google t-shirt.
Judging from the data and search queries,
our mobile user may take two common approaches.
They may perform a navigational query
for Google Store or a transactional query
for Google t-shirt.
In either case, we can start to outline the workflow
and take note of each step.
In the search query and appearance stage,
we can make sure that our site's appearance and search
results is appealing.
The page has a descriptive title and meta description.
Once the user, let's pretend his name is Tom,
selects our page in Search Results,
we're in the page content and engagement stage.
In the workflow with the query Google Store,
once Tom clicks a search result, he comes to our home page,
where he must select his region.
Then he's taken to his region's main shopping page, where
he may select the Wearables category
with the subcategory men's t-shirts.
Once at the men's t-shirt category page, hopefully,
Tom will click through to an individual product page, where
he can add to cart and, eventually, get
to the conversion stage and make a purchase.
By the end, Tom has both completed his task
and helped our business.
Thanks, Tom.
Most real users workflow has many more twists
and turns than this.
Still it's useful to put yourself in your user's shoes,
like Tom's, to get a sense of what
it takes for task completion.
Now, to prioritize improvements in the workflow,
first eliminate any unnecessary pages,
check for extraneous steps that can be merged or removed
in both the site engagement and the conversion stage.
Second, maintain information sent.
Third, fix the remaining pages in the workflow
that show poor user experience metrics.
Fourth, minimize user interaction cost
throughout the task.
That's interactions such as typing, clicks, expands,
and scrolls, which are particularly
tedious and error-prone on smartphones.
With the first action, reducing the number of pages required
in the workflow, I can test removing the homepage map
and displaying based on the user the appropriate main shopping
page.
This may also help reduce the over 50%
homepage drop-off we saw earlier.
We just need to make sure to provide that tourist
experience.
To maintain information sent, we'll
strive to make the navigation and the calls
to action intuitive.
Remember the questions we asked while researching
the mobile users intent?
Now, it's time to test those ideas.
Since the top query term in on-site search was shirt,
that may mean that users either didn't notice or didn't
associate the category Wearables with t-shirts.
That could signal a loss of information sent.
We could test whether Wearables is a helpful category
name for t-shirts and whether moving the category to one
of the earlier options in the list
increases micro-conversion for this task.
Fixing pages with poor user experience metrics
was actually filmed as a separate video.
If you're unable to watch it, areas to focus for improvement
are performance, usability, and content.
Once that's complete, look to minimize the interaction
cost required for completion.
In other words, create a workflow that reduces typing.
Perhaps use auto-complete, previously
entered values, or auto-advancing fields.
To simplify checkout, you can offer third-party checkout
options.
One click to delete pre-populated values
also saves interaction cost, rather than requiring the user
to repeatedly hit Delete for each character.
And while in Delete mode, fully clear the Input field.
We found this issue as we audited the Google Store
workflow.
When trying to add to cart, the item count
keeps a zero displayed even though it should be clear.
This can make users feel they have
to delete the phantom number.
Or if they add one item, the number will append and make 10.
It's easy to become over familiar with your site,
so definitely check the first-time user experience
and have new users perform common tasks.
You may learn to improve usability
with things like password and clear text, option for users
having trouble logging in, as well as errors that are clearly
highlighted and auto-focused for faster correction.
And with that, we've concluded the third and final section
for optimizing the top mobile tasks on your site.
It's my hope that you now feel more
empowered to make improvements.
Thanks for watching.