Social Web Tips

by Paul Jahn

It was just over a year ago when I chatted with Hanan Lifshitz of Palore
for the first time. We’ve spoken a number of times since then, and he
and the Palore team have done some great things in providing local
business data. I’m very happy that Hanan has agreed to do an online
interview.

Paul: First of all Hanan, thanks for taking the time. It’s been awhile since we have spoken. In a nutshell, what have you and Palore been up to and can you give the Search Engine Guide viewers an overview of your services?

Hanan: Paul, as always it’s a pleasure speaking with you. Palore is a data provider in the local search space. We collect information about millions of SMB’s from hundreds of local search sites, IYP’s, vertical sites and news sites. After aggregating and normalizing this data, we syndicate it to companies that need access to large volumes of rich and accurate local business information, quickly and inexpensively.

Our data services range from simple collection of local business data that appears on government sites, to advanced analysis of ad spending by SMB’s online. Needless to say, all the data is collected from sites that allow scraping through their robots.txt file and in line with US copyright laws. 

Paul: You provide a couple of different services. One of them is content for site enrichment and the other is content for sales leads. Can you explain the difference between the two?

Hanan: Initially, we provided a simple web data extraction service to companies that preferred to outsource this activity instead of dealing with the technical challenges in-house. Typically, these companies wanted to enrich their sites’ listings with information such as opening hours, reviews and price, in order to improve their user experience and SEO.

In recent months we were approached by a number of SEM, SEO and IYP companies that wanted to use our web extraction capabilities to determine which local businesses are advertising online. This is how it works - as we process millions of local records, our system analyzes the different characteristics of each business listing, such as background color and logo size, to determine whether it is a free or paid listing. We take into account the specific advertising program offered by each site, their pricing model and other relevant data such as which sites are affiliated. The result is an accurate report of which local businesses advertise online, how much they spend and on which sites, on a nationwide, multi-vertical base. Our clients then use this data to improve their sales forces’ productivity and to conduct competitive analysis.
 
Our clients have found that it is much easier to sell services such as SEM to landscapers and roofers who already advertise online than to convince someone who has been advertising only on print YP over the last 20 years and who doesn’t understand what’s wrong with that…

You can see some of our high level data reports about local business ad spending on our Palore blog.

Paul: Since I last spoke with you, Palore transitioned from a consumer service to a data syndication service. Can you explain this transition?

Hanan: Indeed, we started with a consumer application that used our local business data. The application had a good uptake with some 150,000 downloads during the first few months, but then we were approached by a number of companies that wanted to buy our data. Initially we refused to give it away, but when we were told by one of the largest local search sites that we have the most comprehensive database in the world about local businesses, and that they were willing to pay handsomely for that data, we decided to make the transition into becoming data providers. I think this was a good decision since it enabled us to focus on data extraction and really boost our capacity.

Paul: I’ve known Greg Sterling for a quite a few years now and he is one of the people I look up to the most when it comes to local search. He’s also on your advisory board. How did this relationship come into play?

Hanan: We have been reading Greg Sterling’s Screenwerk blog for industry news and analysis ever since we entered the local search space. Since Greg was right across the Bay Bridge from us in San Francisco, we asked if he would meet with us, and he did. Greg was so helpful and friendly in that meeting and in subsequent meetings that we invited him to join our advisory board. Greg has been an endless source of good advice and insights and we feel very fortunate to have him on board!

Paul: I’ve always considered local search as bringing users from their computers to a physical business door. For Palore, what are the main differences between online and offline content sources?

Hanan: Palore is a child of the continuing shift from the offline world to the online world in local search. While traditional content providers have relied on print directories, government records and other offline sources, Palore is a pure online play. This has several implications:

• Freshness - our content is always up to date. For example, in our ad spend service we can provide a list of plumbers who have advertised last week on a hundred different sites. The offline equivalent of that is a list of plumbers who have advertised on a print YP last year…

• Unique content - our content is a reflection of the new content types that are available online. Things like reviews, online reservations, company websites and online ad spending have become very relevant to our clients and they can be collected almost exclusively online.

• Scale - from an operational perspective, collecting online data using crawlers is much more scalable and cost effective than transcribing print pages in Costa Rica…

Paul: Local aggregate data has been pretty huge for companies who provide this information to portals such as Yellow Pages. How does Palore differ from companies like InfoUSA, Acxiom, and Localeze?

Hanan: I think all these companies have done a great job in providing content to this growing industry. As a technological start-up, Palore has taken a different approach of focusing just on online data. While we don’t see ourselves competing with these companies on base data such as name, phone number and address, I think we can complement their data with new information that is available only online (see my previous answer).

Paul: I often consider myself technologically inept, which is strange since I make my living on the Web. What does it take to collect information from literally hundreds of sites?

Hanan: Extracting online content about millions of businesses from hundreds of sites is not an easy task, and has required many man years of R&D work.
 
The process starts with identifying the right (and copyright approved) content sources and deciding what parameters we want to extract from each web page. We then use our crawler infrastructure which consists of dozens of servers, spread throughout the US internet backbone, to access millions of pages from the chosen sites. Reaching this information, also known as the “deep web” (http://en.wikipedia.org/wiki/Deep_Web) requires special techniques such as automated address form filling.
 
Next, each business page is analyzed and the relevant data is extracted from it. We then merge the data from the different sources, reconcile any discrepancies, clean it up, QA it and normalize it to a unified format.

While many web extraction companies deal with similar challenges, the local nature of our industry makes this especially difficult. Many of the hyper local sites were designed using obsolete technologies, with strange implementations of navigation methods, text encodings, cookies and Ajax that are hard to deal with. But this makes it ever more satisfying to complete a complex data aggregation project from multiple sources and to deliver a unified feed that really helps our clients.

Paul: Hanan, thank you very much for taking your time to answer these questions. Local search has come a long way in the last five years and will only continue to grow. We look forward to hearing more about you in the future. Hanan, thank you very much for taking your time to answer these questions. Local search has come a long way in the last five years and will only continue to grow. We look forward to hearing more about you in the future. 

Want more from your web site?
Search Influence can help! Targeted Traffic. Increased Revenue. Results Guaranteed. Customized Internet Marketing you can afford.

by Paul Jahn

It was just over a year ago when I chatted with Hanan Lifshitz of Palore
for the first time. We’ve spoken a number of times since then, and he
and the Palore team have done some great things in providing local
business data. I’m very happy that Hanan has agreed to do an online
interview.

Paul: First of all Hanan, thanks for taking the time. It’s been awhile since we have spoken. In a nutshell, what have you and Palore been up to and can you give the Search Engine Guide viewers an overview of your services?

Hanan: Paul, as always it’s a pleasure speaking with you. Palore is a data provider in the local search space. We collect information about millions of SMB’s from hundreds of local search sites, IYP’s, vertical sites and news sites. After aggregating and normalizing this data, we syndicate it to companies that need access to large volumes of rich and accurate local business information, quickly and inexpensively.

Our data services range from simple collection of local business data that appears on government sites, to advanced analysis of ad spending by SMB’s online. Needless to say, all the data is collected from sites that allow scraping through their robots.txt file and in line with US copyright laws. 

Paul: You provide a couple of different services. One of them is content for site enrichment and the other is content for sales leads. Can you explain the difference between the two?

Hanan: Initially, we provided a simple web data extraction service to companies that preferred to outsource this activity instead of dealing with the technical challenges in-house. Typically, these companies wanted to enrich their sites’ listings with information such as opening hours, reviews and price, in order to improve their user experience and SEO.

In recent months we were approached by a number of SEM, SEO and IYP companies that wanted to use our web extraction capabilities to determine which local businesses are advertising online. This is how it works - as we process millions of local records, our system analyzes the different characteristics of each business listing, such as background color and logo size, to determine whether it is a free or paid listing. We take into account the specific advertising program offered by each site, their pricing model and other relevant data such as which sites are affiliated. The result is an accurate report of which local businesses advertise online, how much they spend and on which sites, on a nationwide, multi-vertical base. Our clients then use this data to improve their sales forces’ productivity and to conduct competitive analysis.
 
Our clients have found that it is much easier to sell services such as SEM to landscapers and roofers who already advertise online than to convince someone who has been advertising only on print YP over the last 20 years and who doesn’t understand what’s wrong with that…

You can see some of our high level data reports about local business ad spending on our Palore blog.

Paul: Since I last spoke with you, Palore transitioned from a consumer service to a data syndication service. Can you explain this transition?

Hanan: Indeed, we started with a consumer application that used our local business data. The application had a good uptake with some 150,000 downloads during the first few months, but then we were approached by a number of companies that wanted to buy our data. Initially we refused to give it away, but when we were told by one of the largest local search sites that we have the most comprehensive database in the world about local businesses, and that they were willing to pay handsomely for that data, we decided to make the transition into becoming data providers. I think this was a good decision since it enabled us to focus on data extraction and really boost our capacity.

Paul: I’ve known Greg Sterling for a quite a few years now and he is one of the people I look up to the most when it comes to local search. He’s also on your advisory board. How did this relationship come into play?

Hanan: We have been reading Greg Sterling’s Screenwerk blog for industry news and analysis ever since we entered the local search space. Since Greg was right across the Bay Bridge from us in San Francisco, we asked if he would meet with us, and he did. Greg was so helpful and friendly in that meeting and in subsequent meetings that we invited him to join our advisory board. Greg has been an endless source of good advice and insights and we feel very fortunate to have him on board!

Paul: I’ve always considered local search as bringing users from their computers to a physical business door. For Palore, what are the main differences between online and offline content sources?

Hanan: Palore is a child of the continuing shift from the offline world to the online world in local search. While traditional content providers have relied on print directories, government records and other offline sources, Palore is a pure online play. This has several implications:

• Freshness - our content is always up to date. For example, in our ad spend service we can provide a list of plumbers who have advertised last week on a hundred different sites. The offline equivalent of that is a list of plumbers who have advertised on a print YP last year…

• Unique content - our content is a reflection of the new content types that are available online. Things like reviews, online reservations, company websites and online ad spending have become very relevant to our clients and they can be collected almost exclusively online.

• Scale - from an operational perspective, collecting online data using crawlers is much more scalable and cost effective than transcribing print pages in Costa Rica…

Paul: Local aggregate data has been pretty huge for companies who provide this information to portals such as Yellow Pages. How does Palore differ from companies like InfoUSA, Acxiom, and Localeze?

Hanan: I think all these companies have done a great job in providing content to this growing industry. As a technological start-up, Palore has taken a different approach of focusing just on online data. While we don’t see ourselves competing with these companies on base data such as name, phone number and address, I think we can complement their data with new information that is available only online (see my previous answer).

Paul: I often consider myself technologically inept, which is strange since I make my living on the Web. What does it take to collect information from literally hundreds of sites?

Hanan: Extracting online content about millions of businesses from hundreds of sites is not an easy task, and has required many man years of R&D work.
 
The process starts with identifying the right (and copyright approved) content sources and deciding what parameters we want to extract from each web page. We then use our crawler infrastructure which consists of dozens of servers, spread throughout the US internet backbone, to access millions of pages from the chosen sites. Reaching this information, also known as the “deep web” (http://en.wikipedia.org/wiki/Deep_Web) requires special techniques such as automated address form filling.
 
Next, each business page is analyzed and the relevant data is extracted from it. We then merge the data from the different sources, reconcile any discrepancies, clean it up, QA it and normalize it to a unified format.

While many web extraction companies deal with similar challenges, the local nature of our industry makes this especially difficult. Many of the hyper local sites were designed using obsolete technologies, with strange implementations of navigation methods, text encodings, cookies and Ajax that are hard to deal with. But this makes it ever more satisfying to complete a complex data aggregation project from multiple sources and to deliver a unified feed that really helps our clients.

Paul: Hanan, thank you very much for taking your time to answer these questions. Local search has come a long way in the last five years and will only continue to grow. We look forward to hearing more about you in the future. Hanan, thank you very much for taking your time to answer these questions. Local search has come a long way in the last five years and will only continue to grow. We look forward to hearing more about you in the future. 

Want more from your web site?
Search Influence can help! Targeted Traffic. Increased Revenue. Results Guaranteed. Customized Internet Marketing you can afford.

Facebook is building out its infrastructure so it can keep up with the 250,000 new users it adds each day. In an interview Jonathan Heiliger, vice president of technical operations at Facebook, talks about server purchasing strategy.
In a CIO Sessions interview with Dan Farber (see Dan’s transcript), Heiliger acknowledged the scaling up challenge and […]

by Mack Collier

Think all you have to do is create great content, and readers will beat
a path to your blogging front door?  Think again.  One of the best ways
to grow your blog, is by participating on other social sites.

As I’ve talked about before,
I’m a big proponent of growing your blog, by leaving it.  Here’s five tactics you can use to boost your blog:

1 - Use LinkedIn’s Question and Answers section.  If you aren’t already a member of LinkedIn, check out Jennifer’s excellent recap of why you should be.  The Q&A section is a great way for you contribute value to others.  Simply pick the subjects that you are knowledgeable about, and that tie into your business, and answer away!  As you begin answering questions, others will check out your profile, and also any links your profile has, such as your blog/website.  If you are identified by other LinkedIn users as giving good answers, that helps raise your visibility.  And of course, you can also learn from the answers and questions that other members submit!

2 - Networking on Twitter.  Yes I am a big-time Plurk evangelist, but for most people and businesses, Twitter is still where the action is.  Using a search tool such as Summize, you can search for topics that others are talking about that interest you (or, if they are talking ABOUT YOU!).  Once again, your goal is to create value for others.  As with LinkedIn’s Q&A section, look for conversations where you can contribute.  If you own a pet-grooming business, find conversations among Twitter users concerning, for example,  how to properly groom their pets, and help them with any advice you have.  This will likely result in your gaining some friends/followers, as well as new blog readers!

3 - YouTube.  This is something else that I’ve talked about before, but it’s never been easier for a small business to create videos and post them on YouTube.  Once they are posted, you can add your blog’s URL to the page, and YouTube members can leave comments on the videos.  Don’t overlook responding and engaging viewers via the comments section.  They are giving you valuable feedback that you should respect by replying to.

4 - Facebook Groups.  Same idea here as with Twitter and LinkedIn’s Q&A section, look for ways to create value for others.  If your blog is for your home furnishings store, then look for Facebook Groups around similar topics.  Once you start finding groups that cover topics related to your business/industry, then you can start interacting with the members there, as well as contributing to the message boards that most groups have.  Thanks to Deb for this idea!

5 - Other blogs.  This is still my favorite.  In my opinion, the best way to grow your blog’s readership and awareness, is to participate on other blogs.  Read blogs that are in your industry and that cover the same topics your business blog does.  Comment as you can, and by spurning on the conversation.  This increases the liklihood that other readers and the blogger will click back and check out your blog.  Also, as you find interesting posts on other blogs, create a post on YOUR blog linking to these posts.  This is a great way to share valuable content with your readers, and again, it gives other bloggers a reason to check out your blog.

But the common thread running through these tips is; find ways to create value for others.  It can be by answering a question, by adding a comment, or sharing a link.  The beauty of social media is that as you use these social sites and tools to create value for others, the more value comes back to you.  The best way to grow your blog, continues to be by leaving it.

BTW if you want some more ideas, here’s where I asked for feedback about this article from my Plurk friends, there’s some additional ideas here.

Want more from your web site?
Search Influence can help! Targeted Traffic. Increased Revenue. Results Guaranteed. Customized Internet Marketing you can afford.

by Stoney deGeyter

Here is another round of questions submitted to me during a Website Architecture webinar I gave a couple months back. I wasn’t able to answer most of these questions before or during the presentation so I’ve been answering them in this Q&A series. You can check out Parts one, two, three, four, five , six and seven. This post covers questions regarding session IDs, repetitive vs. duplicate content, robots.txt files, navigation text, and maintaining link juice after a site re-design. Let’s get to it.

If someone cuts and pastes a session ID and emails to someone, do they pick up from the original session or start a new one?
– Kate

That’s a great question and one I didn’t know the answer to… until just now. I put this to the test and came to a definitive conclusion, but that conclusion is only as good as the site I was working on. I suppose it’s entirely possible for every site, or CMS system, to at differently. But we’ll go ahead and conclude that this test is indicative of most sites that use session IDs, and hope it holds true for all. (If someone has a more definitive answer on this I’d love to hear it.)

So here’s what I did. In FireFox I found a site that I knew uses session ID. I copied one of the URLs that carried a session ID in it, and then proceeded to add a product to the shopping cart. I then opened up another browser (Internet Explorer) and pasted the URL in the address bar.

At first glance, the product I added to the shopping cart was still there. But, as soon as I clicked a link I was given a completely new session ID and the shopping cart was cleared.

I then tried to do the reverse, copy the session ID from Internet Explorer, added a product to the cart and then pasted the URL over in my session in FireFox. Because I had already established my own session in FireFox the new session ID has no effect. My original product remained in the shopping, not the one I added over in Internet Explorer.

So to answer your question, it appears that sending a session ID in a link will not give that other person your session, other than briefly and before they click any links.

In reference to repetitive content through a site, is it ok to have the company name as part of the page title on every webpage throughout the site?
– Kirk

I’d say that there is a clear distinction to the search engines from repetitive content and duplicate content. It’s natural to be repetitive about many things. Some clear examples are company names (not just in title tags,) taglines, special offers, navigational element, calls to action, etc.

It’s actually pretty common to include your company name in all your title tags. It’s great for branding purposes. Most SEOs will tell you to put your company name at the end, rather than front, of the title tag, but there is no reason why it can’t be there. The only problem would be is if you use the same, duplicate, title tag throughout your site.

Just remember that a little repetition can be good, but duplication rarely ever is.

Would not having a robot.txt file hurt ranking of a website?
– Dianne

It shouldn’t. You’re in more danger of having improperly constructed robots.txt file than none at all. But that’s not to say that it won’t at some unknown point in the future. If the search engines ever decide that by not finding a robots.txt file then they should conclude that they are not given permission to spider the site then this will effectively block them from spidering your pages.

One thing to keep in mind is that the search engines almost always look for this file when they visit. By not finding it they are given a page not found error. Uploading a generic, all access, robots.txt file will ensure that they don’t get this error and don’t misinterpret that to mean something it doesn’t.

To create an all-access robots.txt file just use the following:


User-agent: *
Disallow:


How do search engines read the keywords incorporated into CSS navigation bars/main menus or images?
– Sharyn

If you’re using CSS for your navigation then chances are that the navigation links are all pure HTML text. The CSS is only responsible for how that navigation displays. So in this case, the search engines have no issues with reading the navigation link text because it’s there on the page just like everything else.

Ive seen cases where, when using JavaScript for site navigation, that the actual text for the navigation links is in the .js file instead of on the page. This creates a problem for search engines and in these cases the engines are usually not able to grab the text or follow the actual links.

As for images, since the search engines can’t read the images they have to rely on the ALT text in the image tag. For all navigation images you should include the words that are in the image in the ALT attribute as well.

How do you redesign a web site without abandoning the “juice” already attained?
– Scott

The best thing you can do is the ensure that all the file names and page/sub-directory locations remain true to the original version. That means you can’t reorganize your site structure, can’t rename file names, and can’t switch to a different programming language that requires the file extensions to change.

This, of course, isn’t always possible with site re-designs and upgrades so the next best solution is to 301 redirect all your old site pages to the new site pages. After a few months the search engines will start passing the link juice from the old pages through to the new pages and nothing will be lost. Going this route, just be sure to never drop the 301 redirects. As long as someone, somewhere, links to an old page, you want that juice to flow to the new page.

Want more from your web site?
Search Influence can help! Targeted Traffic. Increased Revenue. Results Guaranteed. Customized Internet Marketing you can afford.

by Stoney deGeyter

Here is another round of questions submitted to me during a Website Architecture webinar I gave a couple months back. I wasn’t able to answer most of these questions before or during the presentation so I’ve been answering them in this Q&A series. You can check out Parts one, two, three, four, five , six and seven. This post covers questions regarding session IDs, repetitive vs. duplicate content, robots.txt files, navigation text, and maintaining link juice after a site re-design. Let’s get to it.

If someone cuts and pastes a session ID and emails to someone, do they pick up from the original session or start a new one?
– Kate

That’s a great question and one I didn’t know the answer to… until just now. I put this to the test and came to a definitive conclusion, but that conclusion is only as good as the site I was working on. I suppose it’s entirely possible for every site, or CMS system, to at differently. But we’ll go ahead and conclude that this test is indicative of most sites that use session IDs, and hope it holds true for all. (If someone has a more definitive answer on this I’d love to hear it.)

So here’s what I did. In FireFox I found a site that I knew uses session ID. I copied one of the URLs that carried a session ID in it, and then proceeded to add a product to the shopping cart. I then opened up another browser (Internet Explorer) and pasted the URL in the address bar.

At first glance, the product I added to the shopping cart was still there. But, as soon as I clicked a link I was given a completely new session ID and the shopping cart was cleared.

I then tried to do the reverse, copy the session ID from Internet Explorer, added a product to the cart and then pasted the URL over in my session in FireFox. Because I had already established my own session in FireFox the new session ID has no effect. My original product remained in the shopping, not the one I added over in Internet Explorer.

So to answer your question, it appears that sending a session ID in a link will not give that other person your session, other than briefly and before they click any links.

In reference to repetitive content through a site, is it ok to have the company name as part of the page title on every webpage throughout the site?
– Kirk

I’d say that there is a clear distinction to the search engines from repetitive content and duplicate content. It’s natural to be repetitive about many things. Some clear examples are company names (not just in title tags,) taglines, special offers, navigational element, calls to action, etc.

It’s actually pretty common to include your company name in all your title tags. It’s great for branding purposes. Most SEOs will tell you to put your company name at the end, rather than front, of the title tag, but there is no reason why it can’t be there. The only problem would be is if you use the same, duplicate, title tag throughout your site.

Just remember that a little repetition can be good, but duplication rarely ever is.

Would not having a robot.txt file hurt ranking of a website?
– Dianne

It shouldn’t. You’re in more danger of having improperly constructed robots.txt file than none at all. But that’s not to say that it won’t at some unknown point in the future. If the search engines ever decide that by not finding a robots.txt file then they should conclude that they are not given permission to spider the site then this will effectively block them from spidering your pages.

One thing to keep in mind is that the search engines almost always look for this file when they visit. By not finding it they are given a page not found error. Uploading a generic, all access, robots.txt file will ensure that they don’t get this error and don’t misinterpret that to mean something it doesn’t.

To create an all-access robots.txt file just use the following:


User-agent: *
Disallow:


How do search engines read the keywords incorporated into CSS navigation bars/main menus or images?
– Sharyn

If you’re using CSS for your navigation then chances are that the navigation links are all pure HTML text. The CSS is only responsible for how that navigation displays. So in this case, the search engines have no issues with reading the navigation link text because it’s there on the page just like everything else.

Ive seen cases where, when using JavaScript for site navigation, that the actual text for the navigation links is in the .js file instead of on the page. This creates a problem for search engines and in these cases the engines are usually not able to grab the text or follow the actual links.

As for images, since the search engines can’t read the images they have to rely on the ALT text in the image tag. For all navigation images you should include the words that are in the image in the ALT attribute as well.

How do you redesign a web site without abandoning the “juice” already attained?
– Scott

The best thing you can do is the ensure that all the file names and page/sub-directory locations remain true to the original version. That means you can’t reorganize your site structure, can’t rename file names, and can’t switch to a different programming language that requires the file extensions to change.

This, of course, isn’t always possible with site re-designs and upgrades so the next best solution is to 301 redirect all your old site pages to the new site pages. After a few months the search engines will start passing the link juice from the old pages through to the new pages and nothing will be lost. Going this route, just be sure to never drop the 301 redirects. As long as someone, somewhere, links to an old page, you want that juice to flow to the new page.

Want more from your web site?
Search Influence can help! Targeted Traffic. Increased Revenue. Results Guaranteed. Customized Internet Marketing you can afford.

Content management software vendors are pondering social networking features and are expected to collaborate or acquire so-called white label networks to make their enterprise apps more social.
That’s one of the takeaways from Jeremiah Owyang, an analyst at Forrester.
Owyang’s working theory is that the big enterprise content management software (CMS) vendors–EMC’s Documentum, Interwoven, Vignette, Stellent etc–are […]

Content management software vendors are pondering social networking features and are expected to collaborate or acquire so-called white label networks to make their enterprise apps more social.
That’s one of the takeaways from Jeremiah Owyang, an analyst at Forrester.
Owyang’s working theory is that the big enterprise content management software (CMS) vendors–EMC’s Documentum, Interwoven, Vignette, Stellent etc–are […]

by Manoj Jasra

Want more from your web site?
Search Influence can help! Targeted Traffic. Increased Revenue. Results Guaranteed. Customized Internet Marketing you can afford.

by Manoj Jasra

Want more from your web site?
Search Influence can help! Targeted Traffic. Increased Revenue. Results Guaranteed. Customized Internet Marketing you can afford.