Can non-technical Marketers really benefit from Google Analytics’ Instant Answers?

Author: Florian Perl Pertynski

Google Analytics is launching more and more features for the non-technical marketers. This makes perfect sense as the Google Analytics user groups are becoming increasingly diverse in competence and profile.

In late 2019, Google released a version of Google Analytics called app-+web. This version includes the function ‘Instant Answers’. The intention of Google is to offer a fast track for insights and improve usability, especially for non-technical marketers.  In short, Instant Answers’ offers you ask a wide range of questions about your Analytics data in natural language and get quick answers.

In this blog post, I am going to evaluate, how it performs and give my perspective on how best to use it.

(If you’re already familiar with the idea Instant Answers, scroll down to the More Questions… section – hopefully you’ll see some new stuff there!)

Part of our job as analysts is to deliver insights to marketing, whose job in turn is to turn it into revenue (or savings). This December, GA App+Web offers to do this job for us. Enter Instant Answers!

And, actually, it does a good job at it – in its way:

GA App+Web has been subtly preaching “good practice analytics”, offering us the “engaged” metrics. And so, rather than a count of all users, we get a count of engaged sessions – a refined answer to a possibly over simplistic question.

Interestingly, clicking on the result takes us to a generic report which actually does not contain an answer to the question. So the inline answers seem the best shot – for now.

More questions – more answers

Let’s try a few more:

1. BASICS: count of events, metrics

This is the basic one. Try:

  • sessions
  • users (unfortunately, active users returns just Users)
  • name of event, such as page_view or view_item

2. BASICS: total count of multiple events

Voilà! “And” can be successfully used to combine count of multiple events. Look how neatly it is translated into the logical operator OR.

3. FILTERS: location, time frame

4. FILTERS: multiple locations (sum of event count), time frame

It looks like that the “and” is interpreted literally as a logical operator AND (with any dimensions or metrics, seemingly). In this case, we mean OR. Let’s try to be literal:

Something is definitely happening here, but still we are not getting a result no result. It looks like with dimensions the “or” operator is going to mean juxtaposition, not the OR condition.

5. FILTERS: time ranges

Instant Answers recognises time ranges well. It is all about speed, so the input format doesn’t have to be very strict.
The default range is last week, so if you don’t specify the time range, you will see last week’s data. But you can also try these with the sessions metric:

  • last month, year
  • this month, week, year
  • 2019
  • nov (returns the current year’s November data)
  • nov 2019
  • 11/5/2019 (mm/d/yyyy)
  • 5 nov (returns the current year)
    or even
  • 5-6 nov or 5 nov – 4 dec

6. COMPARISON: day on day, month on month, week on week, year on year

This is really instinctive: just use DoD, WoWMoM or YoY growth (case-insensitive) with a time range (or without, for the current month or week).

Just watch out when you add a dimension filter – such as traffic source: Instant Answers may kick in with Acquired Source rather than Last Click Source (the latter being Source in Universal Analytics):

7. CALCULATIONS: conversion rate

Did I really expect this to work?

I’ve tried:

  • conversion rate
  • conversions per session
  • conversions per user
  • conversions to sessions

with no success. Definitely a wish-list item.

8. TIMELINES: trend graphs

Inline answers are simple text – sometimes embellished with an up or down arrow to mark the result of a comparison. In this case, however, Instant Answers gives an answer in a separate card, accessed with a click on the inline answer:


Query Instant Answers for a share in traffic of a particular traffic segment:

or ask for a full shares report:

Interestingly, you can ask for a table with multiple metrics (in this case, both will show a given segment’s share):

but not multiple dimensions (if querying for shares):


Let’s ask for more reports (rather than inline answers). Querying the top values, it looks like we can ask for multiple dimensions, e.g. 

It seems that when querying for top values with a single metric, where the result contains up to 5 rows, Instant Answers can serve us a bar chart:

The limitations are that you can’t:

  • expand the report (the columns are a bit squashed)
  • share or download it (although in some rare cases there is a Go to report link)
  • edit the report (i.e. use it as a base for an Exploration table or graph)
  • view a graph and table at the same time (which is also a limitation of the Analysis module)

We speak, Google listens

At the end of November, some users (including me) were invited to test the yet unpublished Instant Answers feature. Part of my feedback was that the questions about “traffic from…” were interpreted as geographic rather than traffic source-related. I’m actually impressed that this has changed. So, folks, if you’re ever invited for a test drive, do accept, as product dev teams seem to lend an ear to meaningful feedback.


Instant Answers is a fully functional, flexible and intelligent tool for quick ad-hoc data queries. Working with Instant Answers is a discovery process – with some paths already mapped, I’m sure there are (or there will be) more ways to use it than I’m aware of right now (feel free to add some in the comments!)

Still, I’m not entirely sure it can make to my top 5 list of App+Web priorities – looking from a business-case perspective…

I somehow still think that a non-technical marketer (see Simo Ahava’s classic here) will look for insights with the analyst, rather than Instant Answers. True, the query format is flexible, but is it flexible enough?

An analyst, on the other hand, will probably want to see a full table or graph, then add some filter or segment, drill down, save the data etc. So the Analysis module or BigQuery (or Data Studio) will likely be more handy. Still, this falls under “Intelligence”, so looks like Google are on track with the roadmap presented earlier last year. Let’s see if they keep that course with the next developments.