let's have a session on correlation so when we're thinking about marketing data we will need to interpret that data and if you interpret that data we need to use interpretation methods now the three interpretation methods are correlation confidence intervals and extrapolation so if we can look at correlation in this video then correlation is simply understanding the strength of a relationship understand the strength of relationship and it's likely to be between two things two variables and it's often that they have a cause and effect relationship for example online advertising spend versus website visits and if we
chop up here online advertising spends being the cause and the effect is website visits now we can see here that as online advertising spend has increased the website visits have increased so that's an example of a positive correlation because it's clearly going in this direction but it is extremely strong this is a very strong positive correlation because those dots there are very tight and the line of best fit can move through it perfectly now let's put a look at another correlation so here we've got online advertisement spend and if we compare online advertising spend -
not website visits but website sales we can see that the crusher is marking each of those points well this time they are not so perfect as before but they are you can put a line of best fit for it they are giving a positive correlation but it is not so much a strong positive correlation so matches a weak positive correlation so there is some sort of correlation there but not as much as with the website visits now we can also have other types of correlations so here we've got online advertising spends still that same item
there but again this time we're going to compare it to brick and mortar sales so physical shop sales as opposed to online e-commerce or in commerce sales and we can see that as online advertising revenue increases the amount of sales that I read brick-and-mortar seem to be declining decreasing okay as its decreasing it's providing us with a negative coalition and the lower best fit does almost go through all of those crosses marking each of those points so it is a strong negative correlation now the last type of correlation needs to be aware of is this
time we're looking at online advertisers spend and comparing it to start absenteeism and we can see that the cross is there provide absolutely no correlation whatsoever it's impossible to get a line of best fit through that so that's an example of no correlation now it's worth considering when we're talking about correlation that there can be other things that get in the way and break these relationships up or lessen the actual real cause-and-effect relationship for example external influences is a good place to go so we look at the second one here when we compare online advertising
spend to website sales or maybe it's a weak positive correlation but that's because there's a recession and so there's less consumer income available and they're not willing to spend that income as much and that's why maybe the sounds are produced it's not so much to do with the online advertising spend or in the case of number three maybe the reason why we've got lower brick-and-mortar sales is because actually there's a health crisis and that's preventing people from be able to shop in a physical environment and therefore that's why it's tougher so you could actually have
other things that could impact this relationship so you could clearly always use that as an evaluation point that maybe there's this third and fourth external factor there that's not being represented within the coronation so I hope that helps in correlation I'll see you next step [Music]
