Prioritization Framework: What to Build First

Full transcript

1,099 words · script draft, under review.

Intro

You've finished the gap analysis. You have a list. Forty content gaps, a few missing page types, a weak internal linking structure, and a handful of keywords where a competitor beats you by two positions.

Every item on that list is real. Every item has a score from the last episode. And realistically, you can do five of them this quarter.

So which five? If you just take the top five scores, you'll often pick the wrong ones. A score tells you how big an opportunity is. It doesn't tell you what it costs, what it depends on, or how long it takes before you know whether it worked.

In this episode, I'll show you how I turn a scored gap list into a ranked plan of five items, with a written reason for each. Same example shop as before, and all numbers are made up for illustration.

Why the top score is not the first task

Here's a typical situation on example-shop dot test. The highest-scoring item is a set of comparison pages, one for each pair of popular products. Big demand, winnable results pages, good commercial value. The score in episode forty-three was right to put it on top.

But to build those pages, the shop needs structured product data it doesn't have yet. Someone has to write a template. Then the pages have to be crawled and indexed before they can even start ranking. That's months, not weeks.

Further down the list there's a boring item: rewrite the titles on four category pages that sit on page two. Smaller score. But it takes an afternoon, and you'll see a signal within a few weeks.

The score measured the size of the prize. Prioritization adds everything else: cost, order, and speed of feedback.

Five questions per item

For every item on the list, I ask five questions.

Impact. This is the score from episode forty-three. But I convert it into three bands: high, medium, low. The decimals in a score look precise, and they aren't. Bands stop you from arguing about the difference between point seven two and point seven five.

Effort. Honest working time, including writing, development, review and publishing. Again in bands: under a day, about a week, or a month or more. If you don't know, the answer is "a month or more" until proven otherwise.

Dependencies. What has to exist first? Data, a page template, a hub page that links to the new pages, or a fix to something that blocks indexing.

Time to signal. How soon can you measure whether it worked? A title change on a page that's already indexed shows up after Google recrawls it, usually within a few weeks. A brand new page type has to go through crawl, index and rank first, as we saw in episode seventeen. That can take months.

Confidence. How solid is the evidence? A keyword gap you verified against a live results page is stronger than one that rests only on an estimated ranking database. Low confidence doesn't kill an item. It means you test it small first.

The impact–effort matrix

Now place every item on a simple two-by-two. Impact up, effort to the right.

Top left: quick wins. High or medium impact, low effort. Top right: bets. High impact, high effort. Bottom left: fill-ins, things you do when you have a spare hour. Bottom right: money pits. Low impact, high effort. Those you drop, and you write down why.

Here's where the example shop's seven items land. Rewriting four page-two titles: quick win. Adding internal links from the main hub to twelve buried buying guides: quick win. Fixing two pages that compete for the same query, the cannibalization case from episode twelve: quick win, slightly more work.

The comparison pages: a bet. A size calculator tool: also a bet, and it depends on the same product data.

Thirty blog posts on broad informational head terms looked big on volume. But when we broke those head terms down into buying phrases in episode twenty-five, very little of the volume had buying intent. Low value, lots of work. Money pit.

And a link-building campaign: uncertain impact and high effort. Episode forty-two's light check showed authority isn't what's blocking this shop right now. Parked.

Dependencies and staging

Next, draw the dependencies. This changes the order more than anything else.

The comparison pages need a path for Google to find them. That path is the hub page. So the internal linking fix isn't just a quick win, it's a prerequisite for the bet. Do it first.

And a bet never ships at full scale. I build one template and publish three pages as a pilot. Then I wait until they're indexed and getting their first impressions. Only then do I build the remaining seventeen. If the pilot doesn't get indexed, I've learned something important for the price of three pages instead of twenty.

The ranked plan of five

So here's the plan.

Number one: rewrite the four page-two titles. Fast, cheap, and it also tests your measurement setup before the bigger items arrive.

Number two: the hub links to the buying guides. A quick win, and it unlocks number four.

Number three: fix the cannibalizing pair.

Number four: the comparison page pilot. Three pages, one template, with a clear gate before scaling.

Number five: the size calculator, but only as scoping. Write the spec, check whether the product data exists, estimate the work. Don't build it yet.

Dropped, with reasons on file: the blog posts and the link campaign.

Notice the mix. Three quick wins and one bet in flight, plus one bet being scoped. Quick wins keep momentum and prove that your measurement works. The bet is where the real growth is. A plan with only quick wins stays small. A plan with only bets goes quiet for months, and people lose faith in it.

Where prioritization goes wrong

Three pitfalls.

Fake precision. Ranking by the third decimal of a score. Use bands, and let dependencies decide ties.

Everything is priority one. If five items are all "urgent", nothing is. Force a strict order from one to five.

No written reason. Six months from now, someone will ask why you built the comparison pages before the calculator. If the reason is written next to the item, that's a thirty-second answer instead of a meeting.

And give every item a re-measure date when you rank it. We'll use those dates in episode forty-six.

Wrap-up

Score tells you the size. Effort, dependencies and time to signal tell you the order. The prioritization sheet I use is linked in the description. Next episode: how to write all of this up as a gap analysis report that a client or a boss will actually read. See you there.