SEO Consulting for HUGE Website. How Big Is TOO Big Of A Change?

[BrianGilmore]

SEO Consulting for a HUGE Website. Their h1 tags have instagram/twitter, h2 have their menu/what's trending and h3 is the article title. Here's what I want to do...

MY MAIN QUESTION: This site has tens of thousands of pages, all articles beyond the few dozen category/tag pages they have.

If I change the structure to the following, will it be too much of a system shock to Google? Will this actually HURT them?

Currently on the site: - h1 tags point to Twitter/Instagram sidebar widgets
h2 tags point to the menu/what’s trending section (which is the same on every page)
h3 points to the Title of the Article

I want to change it to this: - h1 tags should delineate the article's name. That's all they should really be used for.
h2-4 should be reserved for article subheadings to be used by the editorial staff.

EDIT: 30% of their >11 million monthly uniques come from search. I don't want to eff with that, but the way that NONE of their pages have optimized words, they have no sitemap, webmaster tools and are still doing this well makes me think that even putting in minimal changes to tidy things up will help them bring it to 70% organic search.

[DougRoberts]

Good advice. When there's a potentially large impact (organic makes up such a large percentage of your traffic) you really do need to tread carefully.

I've seen more than one site that rolled out sweeping changes in, shall we say an overly enthusiastic manner, and accidentally remove themselves from search completely!

I would recommend doing some research and identifying the real low hanging fruit. What queries/topics/categories is there the greatest search opportunity. If you're already doing well for particular terms then there's not much scope for improvement and the impact of getting things wrong is worse.

Can you look at particular pages that are performing badly. Look for landing pages for organic search traffic that have poor engagement metrics. This can identify poorly targeted keywords, or missing/poor content, miss-understood search intent etc.

• Make sure you document everything (with dates!).
• Don't try to do too much too fast. Small steady tests are safest and make sure you give your changes long-enough to see any impact.
• Make sure you have some kind of QA. Run checks before and after you make your changes. It's great if you can have some kind of check list. Watch out for unintended consequences.
• Are you tinkering with the live site or is there a development/deployment process you need to follow are there other people involved? If there is - stick to the process.

[WilliamKammer]

If the traffic is relatively stable on the site, then testing on newly published pages and monitoring their tragectory when compared to previous articles might work.

Also, I think 2 weeks is a bit short. I'd shoot more for 3 weeks to a month if you want to see where they settle. If you're using new pages, just use the same time frame for those as when you compare them to previous article data.

[BrianGilmore]

Thank you!!! That's exactly what I was thinking, but was unsure. How long would you take your sample for, though? Two weeks? They have a super high-impact site, but all their SEO is news-y, so the pages wouldn't have much time to live.

That's kind of my worry about testing, is that their pages are all entertainment news, so they have an expiration date. So declining on page-search that I track may be due to its age and not anything I'm doing.

Anyway, thank you SO MUCH William!

Any other suggestions more than welcome. Thanks again, guys!

[WilliamKammer]

Sounds like the site is big enough that you have the luxury taking a nice little chunk of pages, doing your tests, seeing what happens, and then deciding whether to make the change site-wide. Take a good sample of pages across you site to test on, make sure you know their baseline ranks and traffic to those pages, make the changes, monitor, test some more, etc. This way, no guess work