#that isn’t the one from the wiki I excluded cause you can only sort of see him in the back | Explore Tumblr Posts and Blogs

bombardthehq · 4 years

Text

Observation Reconsidered Jerry Fodor, written 1984, read 9.06.20-18.06.20

fodor's Granny hopes to counter claims that observation is theory-laden

she says: there are two roads to "the fixation of belief" (following Peirce?)

those based on observation those based on inference

there is a 'corresponding taxonomy of beliefs': observational beliefs and inferential beliefs

observation is the more reliable than inference because there are fewer steps to fixing belief; hence, we try to rely on observations (eg. in the sciences)

but some things are inaccessible to observation, eg. ultraviolet light, so inference must be used

observation is important in settling disputes; if we disagree on matters of inference, we can only say we prefer one theory for another. If we disagree on observation, we can simply look and see for ourselves who is right

Fodor takes back the mic...

he wants to defend this view against 'widely endorsed' arguments against drawing a clean theory/observation distinction

so for fodor, observational beliefs are 'theory neutral'

there are three arguements against this view:

ordinary language arguments meaning holism arguments de facto psychological arguments (a fourth 'ontological' argument is excluded as non-realist)

1. THE ORDINARY LANGUAGE ARGUMENT

Fodor says the main contention of the paper is that "there is a theory-neutral observation/inference distinction" & the boundary is set by "fixed, architectural features of an organism's sensory/perceptual psychology"

this is not however how scientists see it / talk about things

the distinction scientists make between observation & inference is 'relativized to the inquiry at hand'

what an experimenter calls an observation depends on the assumptions the experiment makes, eg. response times are taken by measuring a clock, so checking the clock is a way of observing the response times; but if the clock is broken...

this is a concession to an epistemic fact: not every element of the experiment can be tested, they rely on things which are assumed to work in a certain way, such as the clock- the tests of these assumptions have happened elsewhere

however, even if they use it this way, this doesnt 'settle the case against Granny'

she argues that while actual experimenting scientists observations are theory-relative; but there is still a theory/inference distinction outside of 'ordinary language' that no lingustic considerations can decide

ie. we have to do more than socio-linguistic analysis of working scientists [why? where does the theory-inference distinction ever come up outside of actual observations?]

2. ARGUMENTS FROM MEANING HOLISM

he outlines a complicated picture of a graph which represents a theory with each node its entailiments etc. that I dont need to reproduce

what happens is that if one presupposition is displaced, the whole graph distorts

so the 'minimal context' of the meaning of any theoretical postulate is the *whole theory*

he says this argument has done a lot of work for skeptical philosophers since Quine's two dogmas

its possible to accept this holism (which 'granny and I do not') while defending an observation/inference distinction, eg. an epistemic rather than semantic distinction - every statement gets its meaning from its theoretical context, but some depend on more empirical confirmation than others; Quine takes this view

but the observations can never be theory neutral, because what your "observation sentences" mean depends on what your theory is

he gives Paul Churchland as one defender of this view: "[observation terms] position in semantic space appears to be determined by the network of sentences contaning them accepted by the speakers who use them" - sensation itself cant determine the meaing of an 'observation term', only 'networks of belief'

"as we saw, we are left with networks of belief as the bearers or de- terminants of understanding . . . (p. 13). . . a child's initial (stimulus-response) use of, say, 'white' as a re- sponse to the familiar kind of sensation, provides that term with no semantic identity. It acquires a semantic identity as, and only as, it comes to figure in a network of beliefs and a correlative pattern of inferences. Depending on what that acquired network happens to be, that term could come to mean white or hot . . ., or an infinity of other things (14). " (churchland)

so an observation sentence could mean anything depending on theoretical context

Fodor claims this means that “anything might be an observation sentence depending on theoretical context or, material mode, that anything might be observed depending upon theoretical context."

this means you can change your observational capacities by changing your theories

for Fodor this already means that for meaning holism there cannot be a class of beliefs that must be inferential regardless of what theories the believer espouses, because they could always become given meaning by a theory which renders them observational statements [I'm not really sure about this line of argument; why can inference statements always become observations? or: isn't it simply that if I say 'white' and it means a colour, and someone says 'white' and it means a temperature, we have said different things? that the same words, mere words, might refer in one context to an observation and in another to an inference doesn't seem that interesting to me]

Fodor says that recent causal semantic theories may be indicating that holism is not true; not all meanings are dependant on their theoretical context

some of their semantic properties are dependant on their reationship to the world empirically [no footnote! excuse me??? where can I read more about this??]

so Churchland argues: if statements do not get their meanings from their connections with sensations, they must get them from a theoretical context. But neither are the case.

so 'white' doesnt refer to the colour of sensations but the colour of objects; but 'the referntial roles of colour terms tend to be isomorphic' (uh...?) & dont have 'functional roles' (pls explain???) -- in fact, 'white' gets its meanig from its association 'with white things' [sounds like Suassure!]

this however doesnt show that there is a viable theory-neutral observation/inference distinction; only that meaning holism doesnt pose a challenge to it because meaning holism isnt true...

3. PSYCHOLOGICAL ARGUMENTS

the psychological arugment is that there is no distinction to be made between perception and cognition

"perception involves a kind of problem-solving - a kind of intelligence" (Gregory 1970, p. 30) [the wiki article for 'perception' mentions this, ie. how we 'fill in the details' etc]

perception is therefore the process by which we assign probable causes to the vague stimulus we receive

every 'proximal stimulation' is compatible with a great number of 'distal causes' [ultimate causes], so there are many possible worlds where many different patterns would appear & be picked up as the same visual to an observer [Fodor consistently refers to the perceiver as 'an organism' which is a little crazy to me]

perception involves 'betting on the most probably interpretation of sensory data' (Gregory), ie. perception itself involves inferences - "what mediates perception is an inference from effects to causes"

this seems to be true; so psychology provides a direct argument that observations are always inferential, ie. not theory neutral

so if the world is something like a Neckar cube - all ambiguous stimulus, could always be interpreted differently - how can we ever say an effect has this univocal cause? (without inferring it) - that is, if perception involves problem solving, how are probems ever solved?

the answer is that we simply do infer it - from past information... [we're hitting on the problem Merleau-Ponty brings up here - use of memory, etc.]

so sensory information is interpreted according to the observer's 'background theories'

if percepton is usually true of something that is because our theories about it are true

following a Holmes metaphor, "many projections, if you like, of possible criminals onto actual clues." ... "The clues underdetermine the criminal, but the clues plus background knowledge may be univocal up to a very high order of probability." The trick is having the right background information.

he mentions Kuhn & Goodman in connection w/ this argument

Fodor argues however that the evidence doesnt really say this

while perception is interpretive & contextual, it is also 'bullheaded & recalcitrant'

he explains the Muller-Lyre illusion: the <--> line is seen as shoter because its interpreted as further away and >--< larger because its interpreted as closer; we apply size constancy as though looking at objects at different distances & hence the lines seem to us different sizes (ie. we think we're looking at something 3 dimensional) - our interpretation misguides us; it is highly attested but less often in children, who have less time to develop 3d perception

this is usually cited as an example of how background information affects our perception; the background information here is a complex understanding of 3d objects and their 2d projections

Fodor's challenge: everyone by now knows about the Muller-Lyre illusion. Why isnt our perception penetrated by THAT background theory? Why doesnt knowing the lines are the same length make it look like they are?

In fact, why doesnt knowing a drawing is 2d prevent us from perceiving it as a 3d projection in the first place?

This makes it seem that how the world looks to us is unaffected by cognition

Fodor notes that this is true of many things in this sort of experimental data, not just this illusion

[My thought here is that certain cognitions might be more immediate or work faster, eg. spatial reasoning intervenes before memory]

In a footnote Fodor mentions that Jerome Bruner brings up this point & these examples but writes it off quickly - too quickly for him. He considers the problem w/ 'New Look psychological theorizing' to be failing to distinguish 'how much of what you know actually does affect the way you see'

so there are two conflicting facts here: a known plasticity of perception & a known implasticity of perception. The question is, how can a theory of perception accomodate the existence of both?

[Fodor's argument from here on out will probably rest on the strength of this argument for implasticity. I wonder if there's been any more enlightenment on this from perception science in the last few decades?]

Fodor argues this: to get from a 'cognitivist' view of perception to an argument that observations are theory-dependant you have to argue not just that perception is problem solving engaging background information but that perception has access to ALL the background information (or "arbitrarily much" bg information)

So there are really two questions: whether perception is problem solving/inferential, and wether its "comprehensively pentrated by background beliefs (ie. whether it can be theory-neutral)"

So Fodor wants to arrive at a via media between Granny and Bruner: observation IS inference, but that "there is a radical isolation of how things look from the effects of much of what one believes"

ie. observation is inference, but this inferential-observation IS THEORY NEUTRAL

so it doesn't follow that scientists who accept differnet theories observe phenomena differetly [this argument makes sense to me & answers our hesitation about different types of reasoning; its 'theory neutral' in the sense of 'a scientific theory' or some such]

what might psychology of perception look like if perception is both inferential & theory neutral?

perceptual psychology makes a few distinctons:

between perception and cognition - which they treat as contiguous between perception and sensation - which they treat as strongly divided

sensations are the mere stuff that we actually lay out eyes on, which cognitive processes - the perceptual processes foremost - set on to organize ('assign probable distal causes to')

sensation is noninferential & responds exclusively to the Outside perception is 'both inferential and responsive to the perceivers background theories'

perception can be inferential because my background theories supply the premises for inference; sensation doesnt have access to my theories & cant be inferential

so if you want to 'split the difference between Granny & New Look' you need to postulate a 'tertium quid' -- a psychological mechanism 'which is both encapsulated (like sensation) and inferential (like cognition)': the contradiction between inference and encapsulation [which I take to mean 'access to the Outside'] is resolved by assuming our perceptions access to background mechanism is 'sharply delimited'

he says he develops this properly in Fodor 1983

but to briefly treat what these 'modular' perceptual mechanisms might be like:

perception might assign sentence tokens to sentence types

two reasons this is plausible: "its obviously true" (lol), and... in order to understand what someone says you ned to know the 'form of words' uttered, and 'assigning an utterance to a form of words is to assign a toekn to a type' [I'll take your word for it Jerry]

lets assume theres a psychological mechanism for this -- a 'parser'

the parser takes 'sensory representations' of utterances and makes 'representations of sentence types' (so: acoustic representations in, 'lingustic structural descriptions' out)

the assumption that there are 'psychological mechanisms' and that if they do exist they are 'functions from one form of representation to another' is not given, but it reflects the current psychological theories that we've been dealing with

according to empirical evidence he 'wont bother us with' parsing has 'all the properties that make' psychologists say perception is inferential [this actually seems to be a transcription of a lecture, hence no citations]

the 'acoustic character' of an utterance underdetermines its 'structural description', ie. nothing about the sound of the word 'mankind' should make us anticipate that it is a noun, we can only infer that its a noun with refernece to a background theory about language

so as a module... a parser for L contains 'a grammar of L', ie. noun - it then infers from the acoustic proprties of a token 'a characterization of the distal causes of the token', such as the speakers intention, ie. that the speaker uses it to name something...

so the information it can use in its inferences are: information about the acoustics of the token, and information about lingustic types in L 'and internally represended grammar' -- and nothing else. This 'closure condition' makes the parser modular

the New Look parser, contrarily, can bring in any information that the perceiver has

[we arrive a little at my proposal earlier, that the spatial theory applies 'faster' than other types; Fodor instead believes that the parser's access to anything but (in that case) spatial theories is limited by something intrinsic to the parser itself. It occurs to me that we could decide on this empirially: if you could stop someone from accessing their spatial reasoning, via some kind of 'Russian Sleep Experiment', would they begin accessing other bg theories and update their vision to it, or would everything just appear flat?]

The New Look parser would therefore not see anything it doesnt expect to see; it could never tolerate new information, because it would always turn every sensation into something consistent with its background theories about the world.

what the Modular parser means is that people with very different theories can nonetheless see the world in the same way, as long as their parser doesnt have access to theories on which they disagree

secondly, it means that there can be theory neutral observation language ("much current opinion to the contrary notwithstanding") -- so there is a sense where some terms (like 'red') are observational while others (like 'proton') arent.

his gambit here -- and it is a gambit -- is that only the properties of an object that are explained by the theories that the parser accesses are observational, ie. when I look at something red, my colour parser can access a theory of redness and infer things about it, so I am observing its redness; but if I say that its reflecting(?) electromagnetic radiation, I'm not making an observation, I'm inferring from a theory, because my colour parser does not have access to information about electromgnetism & so cannot observe it

[an interesting manouever, Fodor rescues a scientific realism from Kuhn etc by displacing science even further than them]

giving more cases: the parser has access to certain observations about an utterance, eg. 'an utterance of a sentence' ... 'an utterance of a sentence that contains a word that refers to trees', etc., - these are observations bc they are properties the parser can access. But it does not observe the property 'being uttered to deceive John' or 'being ill-advised in the context...'

one can distinguish between observable properties and sensory properties in this way: the 'sensory properties of utterances are plausibly all acoustic and almost all inaccessible to consciousness'

[this would seem to make all music inaudible]

his third point: the modular parser is 'synchronically' impenetrable by our bg knowledge, ie. in this instance. So just because I know about the Muller-Lyre when I look at it doesn't dispel the illusion. But might they be 'diachronically' penetrable, ie. I could learn how to incorporate more of my bg information into my parser?

it can - denying this is crazy, because it would mean a child could never actually learn a language, etc.

however this also doesnt have complete access, ie. not any kind of teaching influences my parser

the parser is only diachronically penetrable within strictly ("perhaps endogenously" ie. physiologically) defined limits; we still perceive the sun as moving despite our 'Copernican prejudices', and it may be that no educational paradigm could change this

"In this case, our agreement on the general character of the perceptual world might transcend the par- ticularities of our training and go as deep as our common humanity"

So returning to epistemology, Granny will have to 'give a bit' and "distinguish between observation and the perceptual fixation of belief. It is only for the former that claims for theory neutrality have any plausibility."

While the Muller-Lyre doesnt LOOK differnetly to us, we still dont BELIEVE that the first line is shorter; our background information is accessible to 'the mechanisms of belief fixation'

belief fixation, unlike observation ('the fixation of appearances'), is 'a conservative process', using *everything* you know.

think of it this way: the parser module is proposing perceptual *hypothesis* which are 'couched in a restricted (observational) vocabulary and are predicated on a restricted body of info' - then we compare it to the rest of our bg theory and our belief is 'consequent on' this background theory

what this means is that our perceptual beliefs are observational only in the 'first approximation', because afterward they have to be regarded with respect to our whole background theories. So the modules limits determine 'what would you would believe about the appearances from the appearances alone', which by no means accounts for all of the perceptual beliefs you might fix [I have a hard time thinking of an example for this; maybe looking into a microscope shows me a bunch of shapes in the first approximation and, then, subsequently, I can believe that I saw a microbe; its possible that it doesnt even need to be so curious - do I see the smooth surface of carapace in the first approximation and then believe I saw a cockroach?]

so the theory-neutrality of observation is not 'infallible'

its importat here that our perceptual access to observation doesnt necessarily bear on our fixation of beliefs. While it would seem we always know our own perception of things, it doesnt seem to be true. The explanation for the Muller-Lyre illusion [as of this article] hinges on us seeing them, in the first approximation, as 3d objects, which is not how any of us believe we see them; and this explanation was not accessed by introspection but by experiment etc.

An imaginary reader objects: what is this heavily attenuated version of observation good for? havent you given away everything the proponents of 'theory-ladenness' asked for?

he quotes Hanson, one of the main proponents of theory-ladenness, who objects to saying 'all normal observes see the same things in x, but interpret them differently' as unable to explain controversy in research science

rather, for Fodor, 'given the nondemonstrative character of empirical inference' its no puzzle that 'there should be scientific controversy'

the epistemological problem par excellence is to explain scientific *consensus*

how is it possible that given the degree of underdetermination of theory by data that scientists ever agree, and agree as much as they do?

for Fodor, the consensus is down to the theory-neutrality of observation

because we all see the world the same & 'independent of our theoretical attachments', we can see when our predictions arent working out

We admit, Granny and I do, that working scientists indulge in every conceivable form of fudging, smoothing over, brow beating, false advertising, self-deception, and outright rat painting-all the intellectual ills that flesh is heir to ... Nevertheless, it is perfectly obviously true that scientific observations often turn up unexpected and unwelcome facts, that experiments often fail and are often seen to do so, in short that what scientists observe isn't determined solely, or even largely, by the theories that they endorse ... It's these facts that the theory neutrality of observation allows us to explain."

In the end Fodor endorses a sort of plain-spoken realism: science is true because it describes objective reality, and to say this realists will have to hold onto theory-neutral observation.

0 notes

filipeteimuraz · 6 years

Text

The Advanced SEO Formula That Helped Me Rank For 477,000 Keywords

Can you guess how many keywords I rank for?

Well, you are probably going to say 477,000 because I used that number in the title of this post. 😉

And it’s true, just look at the screenshot from Ahrefs. It shows the number of keywords I rank for.

But what’s crazy is that I am in a super competitive niche… digital marketing.

So, are you wondering how I did it?

Well, it starts with proper keyword research.

See most marketers start their keyword research with tools like SEMrush or Ubersuggest and they type in a keyword like “SEO”. You then get a list back with hundreds of keyword suggestions with cost per click and competition data.

And once you have a list of keywords you like, you probably do what most marketers do, which is to start inserting them into your website or creating content around the keywords.

Does this process sound familiar?

Well, of course, it does because that’s what everyone has been teaching you to do.

But what’s wrong with this?

This process is like gambling… there’s no guarantee that you’ll rank for these new keywords. And even worse, those keywords may not generate you any leads, sales, or revenue.

But thankfully, I have a process for you that will not only help you rank for thousands of keywords, but it will also ensure that this new-found traffic converts into leads, sales, and more revenue.

Here’s the 5-step process that helped me rank for 477,000 keywords.

Step #1: Focus on the pages that drive revenue

Going after the right keywords won’t guarantee you success.

If you rank a page that isn’t converting well, you’ll get more traffic, but your revenue won’t go up.

Sure, you can eventually focus on conversion rate optimization and try to fix that over time, but you are better off driving traffic to pages that are already generating you revenue.

If you haven’t setup goal tracking, watch the video below as it will teach you how.

youtube

Assuming you set it correctly, let’s look for the pages that are driving your revenue.

You can see from the image above, I sorted the results by conversions.

You now have a list of pages to focus on. But it isn’t as easy as just picking the top page and going from there.

For example, your top page could be a “check out” page, which, of course, won’t do any good if you rank it higher.

Instead, you should focus on:

Product pages

Service pages

Content pages

Once you have a final list of pages, you’ll want to take those URLs and look them up in your Google Search Console.

Step #2: Log into Google Search Console

Once you’re in Google Search Console, you’ll want to click on, “Search Traffic > Search Analysis”.

This will lead you to a report that looks something like this.

You’ll then want to click on the “Pages” option as it will sort the results by top pages.

At this point, you’ll have to go through your list of pages and find them within Google Search Console.

Once you find one of the pages, click on the URL and then select “Queries” at the top.

This will give you an overview of the specific terms that generate traffic to your high converting pages.

Now let’s download the data in CSV format and open it with Excel.

Once you load it up, it should look something like this.

I want you to first sort the data by impressions. Look for the keywords that are generating the highest impression count as those keywords have the potential to drive the most traffic.

If you feel those keywords are relevant to your product or service that you are offering, make sure you include them within the title tag of your website.

You won’t be able to add all of the keywords to your title tag because it is limited to roughly 60 characters, but adding a few of the most popular terms will ensure that you are going to get higher click-through rates, which will boost your overall search rankings.

Once you’ve adjusted your title tag, let’s do the same with your meta description.

Meta descriptions can be longer these days. Google is ok with roughly 300 characters. So, feel free to sprinkle in a few more keywords, but make sure your meta description still flows in a readable sentence.

And before we get back to the Excel sheet, let’s expand your content by adding in some of the keywords you don’t rank high up on page 1 but should.

You can do this by adding more content to your page, or if you can insert the keywords without “stuffing” them in (just make sure your content flows and provides value).

Now let’s head back over to Excel. You should see a filter icon that looks something like this:

Select column E, as this will select all of the keywords based on their rankings. Then click on the “sort & filter button” and then select “filter”.

You’ll see a table that pops up. Unselect any numbers that are 1, 2, or 3.

You’ll also want to unselect any number that is 11 or greater. This will show you all of the keywords ranking on page 1 that are NOT in position 1, 2 or 3.

These are the keywords that have the most potential as the top 3 positions generate 20.5%, 13.32%, and 13.14% of the clicks respectfully. You want to be in the top 3 spots as that is where the majority of the clicks are happening.

By, having a list of keywords that are in position 4, 5, 6, 7, 8, 9 and 10… you can now focus on moving them up.

You’ve already done the hardest part which is getting on page 1. It’s not that much more work to get into the top 3 spots (at least for most keywords).

You’ll want to take all of the keywords that are relevant to your page and see if you can blend them into your content or headings without ruining the user experience.

This may mean that you’ll have to re-write your content and make it double the length.

Or if you have a product or service page that you are trying to rank higher, it may mean that you can’t include all of the keywords as it will ruin the user experience and hurt your conversion rate (but you can probably include a few more).

Over time you’ll find that your rankings will slowly climb for keywords that will bring in more revenue.

Step #3: Add in related keywords

You know what’s one thing I love about Google? They are really generous when it comes to giving marketers data. From Google Analytics to Google Search Console… Google has some amazing tools!

Another product I love (technically it’s more of a feature) is that Google shows you all of the related keywords to the ones you are already ranking for.

This is going to be a manual grind, but it’s worth it.

Log into Google Search Console and look at the top 10 keywords that you rank for. You can get this data by clicking on “Search Traffic > Search Analytics”.

Make sure you exclude any branded terms and compile a list of your top 10 terms.

Now go to Google, type in each of these keywords manually and scroll to the bottom where it says, “related keywords”.

Google gives you a list of other popular terms that people are typing. What is beautiful about this list, is that these keywords are already related to the one you are ranking for and they are typically much easier to rank for.

So, for neilpatel.com, I already rank for the term “SEO.”

So Google is telling me that in addition to the word “SEO,” people are also searching for:

what is seo and how it works

seo definition

what is seo marketing

how to do seo

seo wiki

seo google

seo tutorial

seo company

As you can see I already integrated some of those phrases to the page on my website that already ranks for SEO.

You should do the same. It’s an easy way to rank for more relevant keywords, boost your traffic, and, eventually, your revenue.

If you do this for your top 10 keywords, you’ll have an additional list of 80 keywords (8 keywords per term).

And by integrating these terms into your site (only when it makes sense, don’t spam) you’ll quickly rank on page 1 for dozens of other terms.

If you want to go crazy like me, you can do this for 1,000 terms, which will then give you suggestions for an additional 80,000 keywords!

But again, don’t force it and ruin your user experience. This will hurt your conversion rate. You should only add keywords when it is natural and makes sense for the user.

And hopefully, you selected keywords from pages that are driving your revenue (remember Step 1!). The last thing you want is to spend time increasing your rankings and find that your revenue isn’t going up.

Step #4: Go after the low-hanging fruit

Have you noticed that there is a huge difference in traffic between the pages on your site that rank on page 2 compared to the content ranking on page 1?

Like most marketers, you probably don’t notice it because your pages that rank on page 2 of Google don’t get much traffic… which causes you to forget about them.

It’s sad but true.

So, let’s fix this!

Log into SEMrush, type in your domain name, and click on “Organic Research > Positions”.

You’ll want to look for all of the terms that you rank number 11 or 12 for.

You can do this by using the filter setting (just copy the settings in the screenshot below).

You’ll have a list of keywords that are almost on page one.

Now just make sure those keywords are pointing to pages that are responsible for driving your sales, leads, and revenue (go back to step 1 if you don’t know how to do this).

For the keywords that aren’t driving sales or leads, you can ignore them for now.

For one the ones pointing to pages that are driving sales or leads, perform a Google search for each of those keywords.

Now compile a list of web pages that are ranking above you.

Take those URLs and plug them into Ahrefs. Once you plug in each URL, click on “Backlinks” in the left navigation bar.

This will show you a list of sites linking to your competition.

I want you to get in touch with each of those sites and beg for a link. Here’s an email template you can use (you’ll have to modify it to fit your site).

Hey [Insert website owners name],

I noticed that you are linking to [insert competing web page] from [insert the URL of on their website linking to the competition]. Did you know that the page you are linking to isn’t the best resource for your website readers?

It’s missing [insert multiple points on what that competing page lacks].

If you want to provide an even better experience for your website readers, you should consider linking to [insert your URL that you want to rank higher] as it has [insert why your web page better than the competition].

Cheers,

[Insert your name]

You’ll find that after emailing hundreds of sites that only 3% to 5% will link back to you assuming your page is comparable to the competition.

If you can’t get at least 3% to link back it means that you either didn’t do a good job modifying the email template or your page sucks compared to your competition. 🙁

I know this is tedious work, but it’s a great way to boost your traffic.

Just think about this stat when doing the link outreach… 91% of searches never go to page 2. Or as my sales team says, page 2 is the perfect place to bury a dead body.

Step #5: Attract buyers before they are ready to buy

Another reason I love Google is that they have this neat tool called Google Correlate.

What Google Correlate does is shows you search patterns. In other words, they show you what your customers are typing in weeks or even months before they become customers.

And if you want to upsell your users, you can use Google Correlate to see what your customers type in weeks or months after they become a customer.

This will help you determine what products or services to offer assuming you want that upsell revenue.

Here’s how it works… let’s say you are selling beard oil. You type in “beard oil” into Google Correlate and you can see what people are typing in before they become customers.

As you can see some of the keywords people are typing in are…

beard oil

beard oils

flannel outfits

trek farley

sweater crop

beard products

best beard oil

sweater crop top

acne studios

what is beard

To get those results I got, I selected “-3” weeks.

I am looking at what people typed in 3 weeks before they searched in beard oil. That’s why I put a “-” sign before the number 3 to see what they typed in before they searched for my main keyword.

If you want to attract more customers and build brand loyalty with people who may be interested in beard oil products, I would create content around the best beard oil or flannel outfits that go well with beards.

That’s what I got from the Google Correlate data.

And if I wanted to figure out what products to create in order to upsell my beard oil customers, I would perform the same search on Google Correlate but use a positive number such as “2 weeks.”

Based on the data above, I would consider offering beard balm as an upsell as there seems to be a strong correlation.

The cool part about Google Correlate is that you can do this for any keyword and sort the results by the country you are targeting.

Conclusion

I know my method of keyword search is a pain in the butt, but it works.

Just think of it this way…

Creating content on new topics is hard because there is no guarantee a new page on your website will rank for competitive terms.

But if you take web pages that already have traction and you improve them using the techniques I described above, it’s a guaranteed way to generate more search traffic.

Now if you want to create content that focuses on new keywords, by all means, you should do so!

I am not saying that creating new content is a bad idea… heck, I do it all the time.

But consider creating new content after you modify your existing pages that are already driving your traffic and sales.

And when you do go after new terms, don’t forget to use Google Correlate as it will help you gain the right type of traffic (plus your competition isn’t doing it).

So, what do you think of my keyword research and SEO process that I used to rank for thousands of keywords?

The post The Advanced SEO Formula That Helped Me Rank For 477,000 Keywords appeared first on Neil Patel.

Read more here - http://review-and-bonuss.blogspot.com/2018/06/the-advanced-seo-formula-that-helped-me.html

0 notes