Key­words, rank­ings, and SEO have gone hand-in-hand since the begin­ning of time. (Well not quite, but it cer­tain­ly feels that way.) As an SEO, this can be a bless­ing and a curse, depend­ing on cir­cum­stances…

Senior man­age­ment, in par­tic­u­lar, can some­times be guilty of judg­ing per­for­mance by search­ing for a few indi­vid­ual key phras­es — and if they’re not hap­py with what they see, SEOs can find them­selves backed into a cor­ner, in spite of wider suc­cess.

Report­ing and KPIs should always dri­ve action, and so when stake­hold­ers see key­word rank­ings which look unfavourable (and they don’t under­stand the nuance of per­son­al­i­sa­tion, local­i­sa­tion, or the role of a giv­en key­word with­in a broad­er strat­e­gy), they are often are often quick to react, demand­ing per­for­mance improve­ments for the term in ques­tion.

It’s the nature of our indus­try that there’s always pres­sure to improve indi­vid­ual key­word rank­ings. But just how fea­si­ble is it to manip­u­late improve­ments for indi­vid­ual key­words?

What’s an SEO to do?

His­tor­i­cal­ly, if we want­ed to improve a rank­ing for a cer­tain term, we would (more than like­ly) start by seek­ing new links, with manip­u­lat­ed anchor text, in order to boost the rel­e­vance of the page we are link­ing to. How­ev­er, in a post-pen­guin world, with poten­tial for man­u­al penal­ties, this becomes a very risky strat­e­gy.

Now, many SEO strate­gies do not focus on improv­ing an indi­vid­ual key­word — or even a par­tic­u­lar sec­tion of a site:

• Often, some of the biggest oppor­tu­ni­ties to improve per­for­mance come from tech­ni­cal opti­mi­sa­tions which aim to make over­all gains and result in improve­ments on every poten­tial key­word.
• Oth­er big wins fre­quent­ly come from cam­paigns, con­tent mar­ket­ing ini­tia­tives and impact­ful media cov­er­age which, sim­i­lar­ly, impact broad pools of rel­e­vant key­words, or a whole web­site.

That’s not to say that no improve­ments can be made to indi­vid­ual key­words. We can always manip­u­late pos­i­tive change in indi­vid­ual terms through tweaks to titles, head­ings and copy, as well as improve the flow of equi­ty to pages and improve­ments in inter­nal link­ing pat­terns, for exam­ple.

How­ev­er these tac­tics often — and quick­ly — reach a point of dimin­ish­ing returns, or you end up chas­ing your tail (pun intend­ed), par­tic­u­lar­ly with larg­er web­sites. Often, the effort doesn’t jus­ti­fy the return, when there are bet­ter and more effec­tive ways to spend our time.

There are oth­er ways we could poten­tial­ly improve indi­vid­ual key­words through a more long-term con­tent mar­ket­ing strat­e­gy. We can attract deep links to the page that our spe­cif­ic key­words rank for. We can ensure that the links we get are from oth­er sites are rel­e­vant or relat­ed to that key­word. Alter­na­tive­ly, we could attract links to our con­tent piece and indi­rect­ly direct the equi­ty via inter­nal anchor text to our tar­get page…

Ulti­mate­ly, though, there is far less guar­an­tee that we will improve the spe­cif­ic key­words we have tar­get­ed.

Which means…

Real­is­ti­cal­ly, we may see some improve­ment in indi­vid­ual key­words with such tech­niques but it’s at the risk of more worth­while pur­suits. Why focus on a sin­gle key­word, at the expense of the long tail? Why try to shoe­horn deep links to improve an indi­vid­ual page, over improv­ing the over­all author­i­ty of a site, which assum­ing a sol­id hier­ar­chy, will poten­tial­ly improve rank­ings of the site as a whole?

By now, my dis­dain for focus­ing on indi­vid­ual key­words is clear. Focus­ing on a sin­gle key­word, no mat­ter how pop­u­lar, resul­tant traf­fic is like­ly min­i­mal in com­par­i­son to the long tail of traf­fic avail­able relat­ing to said key­words. It’s sim­ply not pos­si­ble to try and chase improve­ments for each indi­vid­ual vari­a­tion. Equal­ly, rank­ings vary from hour to hour, from loca­tion to loca­tion, from user to user.

Challenges with reporting

Humans are crea­tures of habit and habits are dif­fi­cult to break down. Rank­ings are easy to under­stand, so are easy to cling on to as a key per­for­mance indi­ca­tor.

What we need, is a bet­ter way of report­ing suc­cess and to edu­cate by num­bers.

In the past, I have used an aver­age rank of a selec­tion of key­words to indi­vid­ual areas of a site to report on per­for­mance, rather than a sin­gle key­word. But this approach is flawed; the aver­age rank could well be brought down by key­words that pro­vide no real busi­ness val­ue. This is entire­ly depen­dent on the key­words you select, of course, but there’s a ten­den­cy to include those that rank well and allow the vol­ume of key­words includ­ed to creep up, as you dis­cov­er new key­words and equal­ly, remove those that have no vol­ume, or sim­ply don’t per­form. When report­ing and analysing per­for­mance, con­sis­ten­cy is key.

Instead of focusing on rankings, what should we focus on?

My pref­er­ence is always to look to traf­fic and rev­enue, as this is what dri­ves busi­ness. Traf­fic and rev­enue can be report­ed by pages, or groups of pages, defined by themes, depend­ing on the size of the site in ques­tion.

Quite sim­ply, you can’t argue with extra traf­fic and rev­enue. How­ev­er, like oth­er met­rics, this one has a flaw. When busi­ness­es report on traf­fic and rev­enue, they fre­quent­ly do so in the con­text of year-on-year com­par­isons. And with Google increas­ing­ly ‘shak­ing up’ its ecosys­tem, things can get messy.

I have per­son­al­ly seen a num­ber of sites show­ing YoY growth, right up until Google’s intro­duc­tion of the 4th ad spot, in late February/March of this year (result­ing in an effec­tive ‑1 drop to every major com­mer­cial key­word). Despite obvi­ous vis­i­bil­i­ty growth and suc­cess, you may not see YoY growth, as the land­scape shifts around you. But that’s anoth­er sto­ry, entire­ly.

With any report­ing, it’s dif­fi­cult to iso­late suc­cess. Nor­mal­i­sa­tion is one method that can be employed. When mak­ing your desired changes, report the change against a nor­malised lev­el of traf­fic and rev­enue. This allows you to see the change with less noise asso­ci­at­ed with it.

In addi­tion to the ever-chang­ing search land­scape, report­ing at a page lev­el may seem a lit­tle ‘black box’ for senior man­agers — but to me, it’s the only way to show improve­ment and prove real val­ue. This is par­tic­u­lar­ly preva­lent when we don’t have detailed key­word per­for­mance data in our ana­lyt­ics pack­ages (in an age of “not pro­vid­ed” key­words).

Here’s the crux of the problem…

I fear rank­ings will always be the thorn in the side of SEOs, as long as we obsess over them and senior man­age­ment remain une­d­u­cat­ed to the big­ger pic­ture.

If rev­enues have increased by 20%, what does it mat­ter if you’ve lost one place for a sin­gle key­word?

KPIs should help you to under­stand per­for­mance, and to dri­ve action — and rank­ings in iso­la­tion do not.

Why try to specif­i­cal­ly improve a sin­gle key­word, when the actions you take could improve hun­dreds, if not thou­sands? If our objec­tive is to help the busi­ness­es which we work with, and for, to be suc­cess­ful, then it’s up to us to edu­cate, and not to spec­u­late!