Welcome back to “More Than Marketing.” I’m your host, Arsham Mirshah. In the past, I’ve brought on experts on various topics for growing an online business. Now, I’m switching up the podcast format.
The new format will deliver the latest and most impactful digital marketing news. The podcast will tell you my reaction, what the news means for you, and what you should do.
Today, I have four topics.
- Microsoft Advertising’s Similar Audiences now available
- BERT, Google’s latest search algorithm to better understand natural language
- The online ad world is murky. A group of companies wants to fix that
- Backlinko analyzed 5.2 million pages for page speed
Transcript:
– Hey, everyone, welcome back to
[00:00:02.03] “More Than Marketing”, I’myour host, Arsham Mirshah. [00:00:04.08] And I’m switching up
the format a little bit. [00:00:07.00] I used to have guests with me, [00:00:08.08] now I’m gonna try a different format [00:00:11.06] where I give you the
latest and most impactful [00:00:15.03] digital marketing news
and tell you my opinion [00:00:19.00] on it, what it means for you,
what you can do about it, [00:00:21.07] and try that as an angle. [00:00:24.00] Might have guests come on, might not. [00:00:25.08] But for now, let’s get into it. [00:00:27.07] So today, I wanna talk
about four different topics. [00:00:30.08] The first of which being Microsoft, [00:00:32.07] Microsoft advertising, Microsoft controls [00:00:35.05] Bing and Yahoo search engines, [00:00:38.02] they have their own ad
network for display as well, [00:00:42.08] but Microsoft has just rolled out [00:00:44.05] product audiences and similar audiences. [00:00:49.07] And that’s now in open beta
as of September of 2019. [00:00:54.04] And it’s exciting because
it’s new technology, [00:00:59.06] it’s new features you can use. [00:01:01.06] It’s kind of about
time, that’s my opinion. [00:01:05.07] Facebook had look a like audiences [00:01:09.07] from I can’t even remember when, [00:01:12.02] ’cause it was so long ago
I can’t even remember, [00:01:13.09] then Google came out
with similar audience, [00:01:17.06] to expand your audience,
then LinkedIn came out, [00:01:20.01] and now Microsoft has bought LinkedIn. [00:01:23.08] So now Bing and Yahoo has that as well. [00:01:28.04] So product audiences is similar
to like Google remarketing, [00:01:34.03] where you do a search for a product [00:01:37.03] and you don’t buy it and then you go [00:01:40.02] about your way on the merry web [00:01:42.02] and you see an ad for that product. [00:01:44.09] And you’re like, whoa
someone was following me. [00:01:46.08] Yeah, so that’s Google doing retargeting [00:01:50.02] or remarketing specific to the
product you were looking at. [00:01:54.07] So product audience is opening beta [00:01:56.09] right in time for the holiday season, [00:01:59.07] and then it’s triggered just
like Google by product ID. [00:02:03.07] So it kinda takes a product
advertiser to set it up. [00:02:08.07] There’s some technology
and some tech behind it. [00:02:11.06] But, you know, it’s pretty
easy to set up, relatively, [00:02:15.06] and it can be really impactful [00:02:17.01] to get people to come back to your site [00:02:19.01] and buy the product they
were looking at, right? [00:02:22.05] Then there’s similar audiences. [00:02:24.04] Similar audiences is like to expand [00:02:26.08] your remarketing audience. [00:02:28.07] So, let’s say you have, you know, [00:02:30.09] 1000 people that come to your site, [00:02:33.05] presumably those 1000 people
are relevant to your business. [00:02:38.00] Well there are tens of thousands of people [00:02:41.05] that are relevant to your business, [00:02:43.03] only 1000 of them are coming to your site. [00:02:45.09] Now what you can do is
you can tell Microsoft, [00:02:48.06] Bing, and Yahoo to say hey now retarget [00:02:51.07] to not only the 1000 people
that came to my site, [00:02:54.02] but also help me expand that to people [00:02:57.05] who haven’t come to my site but look like [00:03:00.00] the audience that is coming
to my site, my visitors. [00:03:04.00] Microsoft is using AI, machine learning, [00:03:06.08] and looking at, you
know, tons of data points [00:03:09.05] and variables to tell you who, or find [00:03:13.08] these similar audiences, right? [00:03:15.08] So similar leverages our
ability to take partner data [00:03:18.09] and extrapolate it across
multiple Microsoft audiences [00:03:21.05] to improve conversion rates in RY. [00:03:24.02] There’s a quote in here
that similar audiences pilot [00:03:27.04] advertisers saw 70% higher
conversions, says Microsoft. [00:03:32.02] Exciting things, definitely
something to try out, [00:03:36.00] you know, look Binging, you know, [00:03:37.09] ads have been really
efficient for some audiences [00:03:42.09] or some of the clients that we work with. [00:03:45.04] Especially those targeting
maybe an older audience [00:03:48.09] or maybe an audience in
BtB, or subcontractor, [00:03:53.07] or government contractors, or just [00:03:57.01] non-tech savvy folks, right? [00:03:59.03] We’ve found very highly efficient cost [00:04:03.00] reposition on Bing and Yahoo. [00:04:06.00] So definitely something
to check out there. [00:04:10.01] Moving on, next, big
announcement from Google, [00:04:14.02] again back in September, they rolled out [00:04:17.08] their biggest search
engine algorithm update [00:04:23.04] since, they say, 10 years ago. [00:04:26.05] So it’s the biggest since 10 years ago, [00:04:28.00] they’ve had several algorithms come out. [00:04:31.04] Penguin, Panda, it’s like
Google’s starting a zoo [00:04:35.00] or something, they name
them all after animals. [00:04:37.02] I don’t know why. [00:04:38.00] This one was not named
after an animal though, [00:04:40.05] this one, well actually, it
might be named after an animal. [00:04:44.08] So, it’s named BERT, B-E-R-T. [00:04:48.02] Sort of like BERT and
Ernie from Sesame Street. [00:04:51.00] I don’t know if BERT is an
animal or not, no he’s not. [00:04:54.02] I’m not up to my Sesame Street
but that doesn’t matter. [00:04:56.09] BERT, if you are super
technical and you want to know [00:05:00.03] what it stands for it’s
bidirectional encoder [00:05:03.00] representations from transformers. [00:05:07.07] Now, let me make that in
layman’s speak for you. [00:05:11.05] So, a transformer is
a word like to or from [00:05:19.00] or for or no, for example,
if you were to search google [00:05:24.04] for parking on a hill with no curb, [00:05:31.09] before BERT, before the algorithm update, [00:05:34.02] you would get a search result, [00:05:36.05] like the first result would be [00:05:38.00] about parking on a hill
with a curb, right? [00:05:42.04] So it would kind of ignore
the word no or otherwise [00:05:44.09] not give it the treatment that it needs. [00:05:48.01] Whereas now bidirectional
encoder representation [00:05:51.08] of transformer takes that no word, [00:05:54.03] that transformer, and sees
what’s around it, right? [00:05:57.08] No curb, okay, curb is the next word, [00:06:00.00] great, I’m going to show
you a different result, [00:06:05.04] one that is about, you
know, parking on a curb [00:06:09.00] without, parking on a hill without a curb. [00:06:13.00] Another example, for
instance, would be like [00:06:16.00] Brazil traveler to USA needs a visa. [00:06:22.08] So before BERT, you would
see like US citizens [00:06:26.04] can travel to Brazil without
the red tape of a visa. [00:06:30.05] Well that’s US citizens going to Brazil, [00:06:33.01] but if you read the query again [00:06:34.06] it’s Brazil traveler to the USA. [00:06:37.08] So now after BERT it’s saying,
you know, Google is taking [00:06:42.00] that to and saying oh Brazil came after, [00:06:45.07] Brazil came before the
to, the transformer, [00:06:50.04] and then USA came after it, [00:06:53.02] so now let me change my search result. [00:06:54.09] The search result is US Embassy consulate [00:06:59.03] in Brazil, so it’s
actually a better result. [00:07:02.05] So what’s this all mean? [00:07:03.05] Basically, look, Google’s
job is, Google’s mission [00:07:06.02] is to organize the world’s
information, right? [00:07:09.03] So when they release an
algorithm, it’s, you know, [00:07:13.00] their aim is to make the
results for your queries better, [00:07:17.09] more accurate, more precise, more specific [00:07:21.03] to what you’re looking for. [00:07:23.07] What can you do about this? [00:07:27.08] Really what you’re gonna do is just [00:07:30.02] continue to create awesome content. [00:07:33.01] Specific content, content
that helps your users, [00:07:36.00] not content for your search engines. [00:07:38.03] Or to game the search engines, right? [00:07:40.01] Gaming the search engine,
search engine optimization, [00:07:43.00] you know, when search engine
optimization first started, [00:07:45.04] I don’t know, 2004 maybe
or really before that, [00:07:50.08] whenever Google came out, that was when [00:07:52.00] the first search engine
optimization started, right? [00:07:55.04] It was really easy to game the system. [00:07:57.03] Right, you’d put white
texts on white backgrounds, [00:07:59.06] stuff the key words, make a tag, [00:08:01.04] and none of that stuff works anymore, [00:08:03.02] in fact it penalizes you. [00:08:05.02] Google’s really smart, so
let’s not game the system. [00:08:09.00] Gaming it only is getting
harder and harder. [00:08:12.02] But search traffic is growing. [00:08:15.09] The number of people that are searching [00:08:18.07] is growing, ’cause more
people come on the internet [00:08:21.05] every day, they use it to
answer their questions, [00:08:25.02] now there’s voice search, with things [00:08:27.01] like Alexa and Google Home. [00:08:29.00] So search traffic is only growing, [00:08:31.07] queries are probably
becoming more specific, [00:08:35.00] I don’t know that for a fact, [00:08:36.01] but that doesn’t really matter. [00:08:37.07] So you have to create awesome content, [00:08:40.03] specific content, to help your
users and you’re gonna win. [00:08:45.01] All right, moving on, the
online ad world is murky. [00:08:50.02] I would agree with that, generally [00:08:51.04] speaking, as an ad agency. [00:08:54.08] And a group of companies wanna fix that. [00:08:57.05] Good, 16 companies are coming together [00:09:00.00] to bring more transparency to online ads. [00:09:03.07] This is lead by the CEO of MediaMath. [00:09:06.00] MediaMath is a leading, maybe the largest, [00:09:09.01] I’m not sure, programmatic
ad platform out there. [00:09:12.04] So this is advertising technology. [00:09:14.08] You know, there’s a lot
of, kind of, blackbox [00:09:20.03] when it comes to ad tech, specifically [00:09:22.09] programmatic, display, realtime bidding, [00:09:26.07] where you know you as an
advertiser don’t really [00:09:32.08] know when you spend a dollar, [00:09:34.05] how much of that dollar
is going to actually [00:09:36.05] buying the media versus all the middlemen [00:09:40.03] in between that, you know, purchase that [00:09:44.05] media for you or place
it or what have you. [00:09:48.05] So, not only that, there’s
also a lot of fraud out there. [00:09:53.06] Fraud, you know, from
smaller ad tech platforms, [00:09:59.01] they could be just, you know, [00:10:01.03] boasting their impressions, just reporting [00:10:03.04] numbers that are not real,
from reporting impressions [00:10:09.05] as just impressions, not
viewable impressions, right? [00:10:12.05] Like if a page, if a web
page loads but your device [00:10:17.00] is only showing the top
of it, the above the fold [00:10:20.01] we call it, well if there’s ads down here [00:10:24.08] at the bottom of the
page that have loaded, [00:10:27.00] the ad platform is gonna say
yeah, that was an impression. [00:10:29.05] But if the user never scrolls
they never saw your ad. [00:10:32.01] So that’s kind of like murky,
maybe borderline fraud, [00:10:35.05] I’m not sure, there’s
even click farms out there [00:10:38.06] where you can just hire, you know, people, [00:10:42.05] usually overseas,
relatively low cost labor [00:10:46.02] to do certain queries or
to go to certain sites, [00:10:50.01] get the retargeting ads
and see the impressions [00:10:52.01] and click on them. [00:10:53.03] It’s just not, it’s not pretty, right? [00:10:55.02] Especially, again, in
the programmatic space. [00:10:58.06] So this kind of coalition
is brought together [00:11:03.02] to add transparency so from the advertiser [00:11:08.05] to the ad, to the ad
agency, to the ad platform. [00:11:14.09] So basically, you possibly move money away [00:11:18.06] from Google and Facebook, the two biggest [00:11:20.04] ad platforms on planet Earth. [00:11:23.00] I don’t know what’s happening on Mars. [00:11:24.04] I haven’t checked, I’ll
ask Elon, he might know. [00:11:27.03] But, yeah, Google and Facebook, they have [00:11:29.00] the biggest inventory,
they have the biggest [00:11:30.07] kind of beta set, they
have the largest audience, [00:11:33.03] and so a lot of advertisers
go there first, right? [00:11:36.04] But this is actually really good news. [00:11:39.03] Really good news for advertisers. [00:11:42.03] It’s good news for ad agencies, you know? [00:11:46.05] We wanna know where our clients money [00:11:49.07] is going so we can better
optimize undoubtedly. [00:11:53.07] And it’s maybe not so good
news for ad platforms. [00:11:57.05] But I think it is in the long term, right? [00:11:59.04] Like I think more and
more dollars are gonna [00:12:01.07] continue to come online into digital ads [00:12:05.01] I think it will continue to grow. [00:12:07.04] All the projections say that’s the case [00:12:09.06] and, you know, it, we
don’t wanna place that [00:12:13.05] we don’t know where the money’s going, [00:12:16.01] ’cause that can only
hurt the growth, right? [00:12:18.08] We know that there’s RY to be had, [00:12:20.07] it’s just let’s optimize
that and let’s be fair. [00:12:24.07] It’s basically what this group is saying. [00:12:27.06] In the group is Oracle,
IBM Watson Advertising, [00:12:31.08] I didn’t even know IBM Watson
Advertising was a thing, [00:12:36.05] Bluekai was bought by Oracle, [00:12:40.01] they were I think the
biggest data platforms [00:12:44.01] so they would sell data
that you would wanna use [00:12:48.04] for targeting on these
programmatic and display platforms. [00:12:51.08] So good news, good news is the ad, [00:12:58.01] this coalition is coming together to add [00:12:59.09] transparency to the ad world and I think [00:13:02.00] that’s a very good thing moving forwards. [00:13:05.01] Last but not least,
possibly the most important [00:13:08.03] as it pertains to your
website is page speed. [00:13:13.05] BackLingo, a very popular
blog SCO blog actually [00:13:18.07] but now it seems that they’re expanding, [00:13:21.02] I think, to more digital marketing. [00:13:23.04] They did a page speed
test where they tested [00:13:25.05] 5.2 million web pages and
they found the following, [00:13:31.02] that a fully load of a web
page is 10.3 seconds on average [00:13:36.04] on the 5.2 million that
they tested on a desktop, [00:13:39.00] 10.3 seconds on a desktop
and 27.3 seconds on mobile. [00:13:46.03] Holy smokes, that’s almost three times [00:13:50.00] longer to load on mobile. [00:13:51.07] And that’s, you know,
likely due to bandwidth, [00:13:54.06] mobile bandwidth being generally slower. [00:13:58.03] But that is alarming, they
also found that page weight [00:14:04.03] is the number one determinant
of the page speed. [00:14:07.06] So page weight is kind of, you know, [00:14:10.04] how much text is on the page, [00:14:13.08] how much, more importantly,
how big the images [00:14:17.06] are that are loading. [00:14:19.00] There’s other things
such as the javascript [00:14:20.07] and CSS files and other files that load [00:14:23.03] to make the web page work
and look the way it looks, [00:14:26.02] but the biggest weight
comes from images, right? [00:14:29.02] So huge images that are uncompressed, [00:14:31.03] they are the number one determinant [00:14:33.03] for how fast or slow a
website is gonna load. [00:14:38.02] They also found that third party scripts [00:14:40.06] significantly slow down
page loading speed. [00:14:43.08] Each third party script
can increase load time [00:14:47.01] by and average of 34 milliseconds. [00:14:50.06] I know what you’re thinking, milliseconds. [00:14:52.08] All right Arsham, milliseconds
is no time, right? [00:14:55.06] Like it, that’s, that
doesn’t matter, but it does. [00:15:00.00] 10 years ago, and don’t worry
I’ll update that for you, [00:15:02.07] 10 years ago Amazon did a
study and they found that [00:15:05.09] for every hundred
millisecond delay in their, [00:15:09.04] in Amazons ecommerce site,
every hundred second delay [00:15:14.09] would cost them 1% of
revenue, that’s huge, right? [00:15:19.09] So if it’s a whole second, right? [00:15:22.08] One hundred milliseconds
times 10 is a whole second [00:15:27.01] you would, no, yes, thousand
milliseconds is a second. [00:15:30.02] They would see 10% drop in
revenue, that’s crazy, right? [00:15:35.06] Now updated that now, in 2017 Octomine [00:15:39.07] did a study that said
every hundred milliseconds [00:15:42.02] delay in a website can
hurt commercial rates [00:15:45.08] by and average of 7%,
so it got 6% worse now, [00:15:50.06] Octomine is reporting,
and that makes sense [00:15:52.03] to me because, now as of I think 2016, [00:15:55.07] mobile connections and
searches are outdoing desktop. [00:16:03.06] In 2016 mobile took over desktop [00:16:05.05] as connections and searches on the web. [00:16:08.03] So, more and more we’re
living in this culture of now, [00:16:11.08] instant, quick, give me
my result, give me my, [00:16:14.05] you know, my purchase or whatever, right? [00:16:18.00] We, I don’t, maybe millennials of today [00:16:24.00] and the future generations who’s gonna be [00:16:25.08] making the purchases, they don’t remember [00:16:27.03] a time with no internet,
let alone, or dial-up, [00:16:31.00] let alone no internet, right? [00:16:32.06] I mean, I remember
dial-up and I was thrilled [00:16:35.00] when we got cable internet,
we could now game. [00:16:37.08] It was awesome right but anyhow
let me read some other stats [00:16:42.02] Google did, released an image that showed [00:16:46.04] that the probability of a bounce increases [00:16:50.06] by how many seconds
your pages load time is. [00:16:55.01] So if it’s one to three
seconds your probability [00:16:58.00] of a bounce is only 32%, not bad. [00:17:00.09] If your page speed is one to
five seconds it becomes 90%, [00:17:06.02] one to six seconds, only one second more, [00:17:08.03] that 90 becomes a 106%,
and then up to 10 seconds, [00:17:12.01] 123% probability of a bounce,
oh it increases by a 123%. [00:17:19.04] So anyhow, page speed is really important. [00:17:21.04] Now what can you do about it? [00:17:23.00] Well, number one you can
test your page speed, [00:17:26.02] there are tools out
there, Google has a tool, [00:17:27.09] I’ll link to it in the show notes here, [00:17:30.03] Pingdom has a tool, the
tools at pingdom.com [00:17:33.05] I’ll link to these in the show notes so [00:17:36.01] you can get to them
directly or you can just [00:17:38.03] Google pingdom page speed test or Google [00:17:40.08] page speed test and the number one result [00:17:42.09] will be the one that you want. [00:17:45.02] That will show you how long
your site or page takes to load [00:17:50.00] and what you can do about it as well. [00:17:52.09] Typically, you’re probably going to need [00:17:54.05] a developer for this but there are tools [00:17:55.07] to help you compress
your images and you know, [00:17:59.06] remove external scripts
that you don’t need. [00:18:02.05] So, you can look out for those. [00:18:05.02] But yeah, this is very important,
it’s kind of a, you know, [00:18:08.05] web developer, techy type
topic, this page speed thing, [00:18:11.09] but it’s, the reason it’s
important for businesses [00:18:14.09] is because it can directly
impact your conversion rate, [00:18:17.04] therefore impacting your bottom line. [00:18:21.02] So that’s it, covered four things today, [00:18:24.02] and I hope you liked this format, [00:18:26.04] we’ll put, in the show notes
we’ll put all the links [00:18:29.00] to the different articles that I covered. [00:18:31.07] Give us some feedback, let me know what [00:18:33.00] you think about this, I want
to keep bringing the news [00:18:36.01] to you because it helps me
stay on top of this industry [00:18:39.00] that’s changing faster than our underwear, [00:18:41.04] and hopefully it helps
you know what’s going on [00:18:44.04] in our world and what you can do about it. [00:18:47.09] So thank you so much for
tuning in, like, subscribe, [00:18:50.09] share this with a friend
who’s in the digital marketing [00:18:52.09] world and we’ll see you next
time on “More Than Marketing”.
Featuring:
Arsham MirshahCEO & Co-Founder
Podcasts Info:
Categories:
Data + TechMarketing
News + Business
Most newsletters suck...
So while we technically have to call this a daily newsletter so people know what it is, it's anything but.
You won't find any 'industry standards' or 'guru best practices' here - only the real stuff that actually moves the needle.