8 Short Steps To Forecast and Estimate SEO ROI…

Author: Gab Goldenberg

… And Based On That Projected ROI, Get Management Buy-In, Set Priorities and Spend Time Wisely.

Note: If you like this post, please link to this page from your blog or share it on StumbleUpon by clicking this button: Stumble Internal Link Building!

Here’s how to project the ROI you can expect on SEO. This way, when your boss asks you to predict next quarter’s SEO numbers to help him with his forecast to analysts, you predictive task will be easy. Better yet, the ROI forecast you make will be defendable.

The detailed explanation on making your forecasts follows after the flow chart.

Flowchart on how to project and/or forecast the ROI on SEO.

1) Pick some short-tail or mid-tail keywords.
There’s no point picking long tail keywords, because the ROI on them individually is small in absolute terms.

2) Check the keyword for commercial intent.

– Microsoft offers a keyword commercial intent estimator tool. Caveat emptor: The tool’s estimates can vary significantly by day, so just take it as another data point, another road sign along the way.

– Correlate that by asking people to do word associations with the given keyword. You can use Amazon’s Mechanical Turk for that. The associations will show the intent behind a given search.

– Compare the Cost-Per-Click (CPC) estimates Google provides you with for each keyword (set on exact match). The higher the CPC, the likelier there is to be commercial value.

Note: For CPCs under $1, this isn’t very meaningful.

3) Find search volume for your target keywords.

Use Google’s KW tool on exact match to find the exact volumes on target keywords.

It’s crucial to use exact match, and not broad match data, because rankings will result in getting you traffic for your exact match keywords and maybe a few variants. It’s best to be conservative.

3.5) Don’t just accept those search volume projections at face value. Double check.

– Run a PPC campaign.

– If you’re not going to run a PPC campaign (eg if the CPCs are too high), you MUST at least correlate the data to other people’s numbers:

  • Use Wikirank, if you see Wikipedia in the top 10.
  • Compare with Aaron’s tool.
  • Ask friends who’ve been in the same vertical about the search volume. Apparently the accuracy of Google’s numbers varies by vertical, so friends who’ve been there can provide valuable insights. (It’s not too difficult to find such friends once you’ve been in this business for a few years.)

4) Find out whether you can rank. Scope out the competition:

  • Look for backlink numbers and quality,
  • Calculate backlinks/referring domains to find out if there are big brands you need to be aware of,
  • Look at their anchor text,
  • Consider their own domain name – an indicator of anchor text past, present and future, as well as financial backing for marketing.

5) Estimate your costs
including domain, hosting, site design + development, links, man hours.

Note: As pointed out by professional search marketer Simon Serrano in the comments below, certain costs are fixed regardless of whether you do SEO or not. So if you’re doing this with an existing site, then you can exclude those costs. But you’ll probably compensate with the man hours etc.

My perspective in suggesting that the domain and hosting costs be included was that of someone starting a new site from scratch, with the mentality, “I’ll rank and sell to search traffic.” For them, the exact match domain might be a big cost that wouldn’t be incurred otherwise, for example.

6) Estimate the traffic attainable from various positions.
See Aaron’s big resource page on the topic and scroll down about 1/3 of the way for the breakdown of Click-Through-Rate (CTR) percentage by position.

Hat tip to my friend Henry Shih of online jeweler Ice.com for coming up with the idea to “make a range of projections, from conservative to optimistic” last summer. Todd Friesen has also suggested this.

Update: Given Universal search results with maps, news, videos, images etc., you may want to review these results downwards to be conservative. As Simon points out in the comments, the CTR data has changed significantly since the time AOL leaked its numbers.

7) Calculate your potential revenue.

If you’re an affiliate: i) Multiply those attainable traffic numbers by your CTR. ii) Multiply again by the merchant’s conversion rate. iii) Finally, multiply by your average commission. Take multiple values for both your own CTR and the merchant’s conversion rate, so you have a range of estimates from conservative to optimistic.

To simplify, if you know what your earnings per click will be (eg recently work in the niche), you can multiply traffic by EPC.

7b) If you’re an adcents publisher: Multiply that by your CTR and EPC if you’re selling CPC ads.

7c) If you’re a retailer / lead buyer: Multiply that by your conversion rate and then by your average if you’re the merchant.

8) Finally, divide your potential revenue by costs. This gives you your SEO ROI – the return on investment for each dollar you put into SEO.

I do this in a spreadsheet for convenience. If you add my rss feed to your reader, you can download a copy of the spreadsheet for your own use.

This has literally saved me thousands of dollars and countless hours by helping me avoid niches that offered a lower ROI for the same time and money investment…

p.s. If you want to know how to prove SEO ROI on past activities, that’s easy. Open your web analytics, and see how many organic visitors on non-brand keywords turned into leads or sales. Alternately, show your rankings increase if you’re not yet in the top 10, and thus not driving traffic just yet.

Found this post useful? Add my feed to your reader for more advanced SEO tips for inhouse SEOs, consultantts and independent webmasters and affiliates! If you want a price estimate for SEO services, you can get one quickly here.

Tags: , ,

Sidebar Story


  1. Brilliant article, truly insightful. All businesses who focus seriously on SEO should read this. Stumbled and tweeted.

    Comment by Barbara Ling, Virtual Coach - July 7, 2009 @ 7:22am
  2. Nice, i've always wanted someone to do a post like this. However, when it comes to steps 5 and 6, I think this is where most people will falter or be inaccurate because: 1)the CTRs that Aaron Wall uses are based on a sampling of keywords from the leaked AOL data. Many companies since then have determined that these CTRs are not accurate across all industries and kw sets. Additionally, SE's have evolved since the AOL leak and have implemented things like Universal search and provide structured results per keyword sets (ie. Bing) which greatly affect CTRs. The website I manage sees a wide range of CTRs for #1 rankings... anywhere from 60%+ for brand terms to less than 1% for highly competitive non-branded core head terms. 2)I don't think estimated costs to rank should include domain, hosting or any other cost which would occur regardless of whether or not SEO was taking place.

    Comment by Simon - July 7, 2009 @ 1:04pm
  3. Thanks Barbara!

    Comment by Gabriel Goldenberg - July 7, 2009 @ 3:41pm
  4. Simon, those are excellent points of constructive criticism! Regarding #1, I hadn't considered that aspect. It's an important caveat on this information, and you're entirely right! Ironically, I should have known, having written about search engines trying to lower the ROI on SEO. I'm going to update the post to highlight this problem. If you'd like a link, I'm happy to credit you for the insight! Email me and let me know! Gab at this domain. 2) It depends if you're building a site as an SEO play or not. My perspective in writing this was that you're starting a site from scratch with the mentality that you're going to rank and make money selling to searchers. Thus exact match domains can be very valuable - but also an important cost to consider since prices are typically 4 figures +. That said, if the domain is already bought and the hosting is going to be able to handle any extra traffic, you're right that those costs should be excluded.

    Comment by Gabriel Goldenberg - July 7, 2009 @ 3:49pm
  5. Excellent post and very useful for the work I am doing this week. Sorted! Thanks :)!

    Comment by Nishan Khednah - July 8, 2009 @ 6:53am
  6. Thanks for submitting to my blog carnival! I just published your submission.

    Comment by Bill McIntosh on Internet Marketing - July 18, 2009 @ 11:26pm
  7. Thanks a bunch Bill!

    Comment by Gabriel Goldenberg - July 20, 2009 @ 3:28pm
  8. This will help every bloggers, site owners and internet marketer out there. Excellent post. I just bookmarked it now. Thanks for sharing.

    Comment by Vic of BusinessAccent.com - July 23, 2009 @ 11:17pm
  9. Gabriel, This is an excellent post, but there are also some other areas to explore - depending on what the revenue model is for the company you're working with. In many cases, the client may already be ranking ok for some terms - just not converting, due to poor usability of the site, or a complicated checkout/booking systems, etc... By recognizing issues such as these early on, asking what their current close or check out completion rate is during those initial meetings, you can often estimate a conservative % increase based on what you think your SEO efforts can do (tying in your keyword data as well). Depending on the client, the estimated ROI can be 2, 5, 10, 20 x whatever the SEO costs are - which is great to know before hand, so that you don't undercharge.

    Comment by David Pavlicko - July 25, 2009 @ 10:37am
  10. Yes this a a good post. Read some of your stuff before on moz. Rand did one somewhere explaining how you can do a ppc campaign and then calulate estimated results from that. Gonna have to go and dig out the figures...

    Comment by Gareth James - July 25, 2009 @ 3:04pm
  11. David, you raise an excellent point. I think I said something similar in my post, but you’re entirely right that conversion rate is crucial in these calculations. If you download my spreadsheet from my rss feed, you’ll see that I use multiple tables for the sake of comparing ROI at different conversion rates.

    Comment by Gabriel Goldenberg - July 26, 2009 @ 12:01am
  12. Feel free to post the link here if you find it. N thanks for stopping by!

    Comment by Gabriel Goldenberg - July 26, 2009 @ 12:01am
  13. Thank you so much for your nice article, at least since it is from 2009, it is showing the improtance of that specific topic! At the last SES in San Francisco where I went to, was one special workshop about it! I blogged also about it under http://ow.ly/1qP8wa cheers... Adirana

    Comment by Adirana - September 4, 2010 @ 6:55am

Leave a Reply