Agree, @Anonymalous, @RobynJ, OP -- I think you are onto something here. IMO think about Credit Karma. Obviously forget the score, but it knows (at least faster than most CMS) when something happens to your CR (at EQ/TU). If you drill through their app/site it will show you exactly what you see on a CR if you were to pull it manually (annualcreditreport, et al) at that point. That alone proves SP=HP info-wise? HP is just the buddy system for the money changers. The algos get busy working AFTER that.
@SouthJamaica I did a quick google search and every search resulted in the same information. Please see below
>
>
>
>
>
>
>
>
>
/
@Anonymalous wrote:
@SouthJamaica wrote:
@Anonymalous wrote:
@SouthJamaica wrote:As I said, the data is different.
In 2 ways.
1. It's taken at a different point in time.
2. The data used to do the prequalification is a limited data set, while the data used to grant or deny the application is broader.
I didn't say anything about algorithms, nor are the algorithms relevant to the question.
If they have access to the same data, but the prequals only look at a limited part of the data set, and the full application looks at a much broader range of the available data, that's two different algorithms.
No it's not. One comes from a search for certain data points. The other looks at the big picture. A search is not an algorithm.
E.g., one might do a soft pull: "FICO 8 score 700-740, no lates, no more than 2 inquiries in 6 months" and get a prospect. But when that prospect applies, it turns out that in addition to the prequal criteria, he has utilization of 40%, owes 90% on several installment loans, and has too little income to support the monthly payments, causing his application to be denied.
... and how do they make the decision, after they search for the data? I think you're one of the best and most knowledgeable posters when it comes to scoring, but you're wrong on this. Your example literally describes an algorithm. Though part of the issue may be that your example seems to describe promotional pulls, i.e. those used for prescreened offers that arrive in the mail. That's not the same as when someone does an online prequal, which involves soft pulling a specific person's credit profile rather than applying a filter to bring up all profiles that match. But even in the case of prescreened offers, it's clear at least some lenders do some processing on the return set, because not everyone is offered the same APR/APR range or SUBs.
What I'd really like to know is whether there is a difference between the data they garner from a soft and a hard pull (no? I don't know?), and why exactly soft pull prequalifications are different from hard pull applications (excluding those based on differences in the underlying profile, which as you noted can be caused by the timing of the pulls). Like I said, there doesn't seem to be a clear reason why they use different decision trees, or what differences there are in the data they're using.
Yes you're right to draw the distinction. I was just thinking of prequalifications which are generated unilaterally by the lender, as opposed to those which are sought by the borrower. I stand corrected.
@SouthJamaica wrote:
@Anonymalous wrote:
@SouthJamaica wrote:
@Anonymalous wrote:
@SouthJamaica wrote:As I said, the data is different.
In 2 ways.
1. It's taken at a different point in time.
2. The data used to do the prequalification is a limited data set, while the data used to grant or deny the application is broader.
I didn't say anything about algorithms, nor are the algorithms relevant to the question.
If they have access to the same data, but the prequals only look at a limited part of the data set, and the full application looks at a much broader range of the available data, that's two different algorithms.
No it's not. One comes from a search for certain data points. The other looks at the big picture. A search is not an algorithm.
E.g., one might do a soft pull: "FICO 8 score 700-740, no lates, no more than 2 inquiries in 6 months" and get a prospect. But when that prospect applies, it turns out that in addition to the prequal criteria, he has utilization of 40%, owes 90% on several installment loans, and has too little income to support the monthly payments, causing his application to be denied.
... and how do they make the decision, after they search for the data? I think you're one of the best and most knowledgeable posters when it comes to scoring, but you're wrong on this. Your example literally describes an algorithm. Though part of the issue may be that your example seems to describe promotional pulls, i.e. those used for prescreened offers that arrive in the mail. That's not the same as when someone does an online prequal, which involves soft pulling a specific person's credit profile rather than applying a filter to bring up all profiles that match. But even in the case of prescreened offers, it's clear at least some lenders do some processing on the return set, because not everyone is offered the same APR/APR range or SUBs.
What I'd really like to know is whether there is a difference between the data they garner from a soft and a hard pull (no? I don't know?), and why exactly soft pull prequalifications are different from hard pull applications (excluding those based on differences in the underlying profile, which as you noted can be caused by the timing of the pulls). Like I said, there doesn't seem to be a clear reason why they use different decision trees, or what differences there are in the data they're using.
Yes you're right to draw the distinction. I was just thinking of prequalifications which are generated unilaterally by the lender, as opposed to those which are sought by the borrower. I stand corrected.
So now that we are all on the same page... What purpose does it serve to tell someone they qualify for the card and auto-decline them on the application literally 2 minutes later? Or the pre-qual saying no offers and you apply anyway and get approved?
seems like a huge waste of time and disruptive to peoples scores for nothing.
Most likely it is become of DTI, Income, length at residence, and one of the big ones believe or not they can't verify your info (something is different on your CR compared to the info you put on the app).
@Anonymalous wrote:
@Vuby22 wrote:Agree, @Anonymalous, @RobynJ, OP -- I think you are onto something here. IMO think about Credit Karma. Obviously forget the score, but it knows (at least faster than most CMS) when something happens to your CR (at EQ/TU). If you drill through their app/site it will show you exactly what you see on a CR if you were to pull it manually (annualcreditreport, et al) at that point. That alone proves SP=HP info-wise? HP is just the buddy system for the money changers. The algos get busy working AFTER that.
Bringing up the credit monitoring systems raises some interesting implications. When we think of someone accessing our scores or reports, we normally think of pulls. But our credit reports aren't full of hourly soft pulls from the credit monitoring systems like CreditKarma, CreditWise, or Experian, and so on. Yet we're sometimes alerted very quickly of changes in our credit reports. That suggests either the monitoring systems continually ping the bureaus to see if there are any changes in the monitored reports, or that the credit bureaus actively push that data out to the monitoring systems. If there is a change, the monitoring systems presumably follow that with a soft pull.
The implication I find interesting is that we know lenders we have a relationship with will periodically pull our credit reports, usually once every month or three. We see this as account review soft pulls, in our credit report. But do the bureaus offer a monitoring service to lenders? Are lenders instantly alerted if there's a change in our profile? Are lenders able to set certain flags (completely hypothetical examples: score drop of 30+ points, utilization passing 80%, new tradeline, etc.), and get alerted shortly after a change is made that matches those criteria?
I'm sure that is a service the big three CRA's they charge for probably exactly that.
I don't see why every CC isn't offering what the Apple Card does, a firm offer based off a soft pull, with a hard pull upon acceptance.
@Anonymalous wrote:
@Vuby22 wrote:Agree, @Anonymalous, @RobynJ, OP -- I think you are onto something here. IMO think about Credit Karma. Obviously forget the score, but it knows (at least faster than most CMS) when something happens to your CR (at EQ/TU). If you drill through their app/site it will show you exactly what you see on a CR if you were to pull it manually (annualcreditreport, et al) at that point. That alone proves SP=HP info-wise? HP is just the buddy system for the money changers. The algos get busy working AFTER that.
Bringing up the credit monitoring systems raises some interesting implications. When we think of someone accessing our scores or reports, we normally think of pulls. But our credit reports aren't full of hourly soft pulls from the credit monitoring systems like CreditKarma, CreditWise, or Experian, and so on. Yet we're sometimes alerted very quickly of changes in our credit reports. That suggests either the monitoring systems continually ping the bureaus to see if there are any changes in the monitored reports, or that the credit bureaus actively push that data out to the monitoring systems. If there is a change, the monitoring systems presumably follow that with a soft pull.
The implication I find interesting is that we know lenders we have a relationship with will periodically pull our credit reports, usually once every month or three. We see this as account review soft pulls, in our credit report. But do the bureaus offer a monitoring service to lenders? Are lenders instantly alerted if there's a change in our profile? Are lenders able to set certain flags (completely hypothetical examples: score drop of 30+ points, utilization passing 80%, new tradeline, etc.), and get alerted shortly after a change is made that matches those criteria?
Yes, the bureaus do offer a service to notify if certain changes happen. They pay extra for certain triggers. Which ties into my next point. You all are assuming the lender is seeing the full report on both the soft, and the hard pull. Pulls cost money, which is why many of the provided free scores from lenders are vantagescore. Lenders can pay less for less data. There isn't two algorithms, it's one algorithm with two amounts of data, one being more complete than the other. Whether or not lenders can pay for just a portion of data for prequal is unknown, but many lenders will use smaller bureaus, like LexisNexis and sagestream in addition to fico or vantagescore. So, while you may have been pre-qualified with just your fico, the official app may turn up something on the additional reports that causes the denial.
