Mime-Version: 1.0 Content-Type: multipart/alternative; boundary="2e42b96d8f1b7cd70d36317e2fc25470063070b987f96e53df900dad4967" Subject: WTF is activation and why should engineers care? From: PostHog To: Hidden Recipient Date: Thu, 6 Feb 2025 18:26:55 +0000 X-Hiring: We are hiring, reach out at header-hacker@emailshot.io X-EmailShot-Signature: Q7BpPInsbpuWt87GOqm5qhrBBnhZhSQ-3ekobKg5_vNOiD8JZk254R96DY7Zhci-3rAZUf51wRlHrvqqTu9ANA== --2e42b96d8f1b7cd70d36317e2fc25470063070b987f96e53df900dad4967 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable View this post on the web at https://newsletter.posthog.com/p/wtf-is-activa= tion-and-why-should Improving activation is one of the most impactful things a product engineer= [ https://substack.com/redirect/69b3c41a-59a4-4587-bcd8-ec08a09a3b75?j=3De= yJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822j-jnN48_jCfgM3m0rbAsF0l24U ] can do, bu= t it=E2=80=99s also weirdly hard to define. People often call it your product=E2=80=99s "aha" moment =E2=80=93 i.e. whe= n a user experiences your product's value for the first time =E2=80=93=C2= =A0but this is vague and subjective. So why isn't there a simple, universal definition? Because: It=E2=80=99s unique to your product. No two products are exactly the same, = so it=E2=80=99s impossible to generalize. This is also why activation bench= marks suck. It=E2=80=99s rarely a single event. You want users to do many actions in yo= ur product: complete onboarding, upload a file, share a link, watch a video= =2E Activation could mean doing any of these, or a combination of them. Activating may mean doing something multiple times. Someone might need to d= o the action multiple times in order to get it. For example, to activate in= to our session replay product means analyzing not one, but five replays. One thing is certain, though. You can=E2=80=99t build a successful product = without nailing it. Read on to understand why. Why engineers should care about activation 1. It's upstream of other metrics In a way, activation is the base rate of your product. Want more revenue? Improve activation. Weak word-of-mouth growth getting yo= u down? Improve activation. It compounds all the metrics downstream of it. If being a product engineer means caring about the whole business (spoiler:= it does [ https://substack.com/redirect/623f13ac-3a04-48da-8443-fd62dde744= 5e?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822j-jnN48_jCfgM3m0rbAsF0l24U ]), = then you should be obsessed with your product=E2=80=99s activation metric. A larger sample size (i.e. more activated users) also makes it easier to an= alyze and experiment on downstream metrics. TL;DR: Everything is easier when activation is strong. 2. It's something engineers can directly influence Activation is part of the AARRR funnel [ https://substack.com/redirect/5a1f= 6ab4-47d8-48dc-81a8-9cf1a16b5277?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822j= -jnN48_jCfgM3m0rbAsF0l24U ]. Above it, at the top of the funnel, is acquisi= tion which is mostly a responsibility of marketers. Below it is retention, revenue, and referral. The last two are mostly platf= orm-level metrics, and normally the responsibility of a smaller growth team= =2E That leaves activation and retention as the responsibilities of product eng= ineers. Activation comes first and directly influences retention, so it=E2= =80=99s a natural starting point. 3. It=E2=80=99s useful across the organization For founders, it=E2=80=99s a signal of product market-fit [ https://substac= k.com/redirect/5cd3d9ab-c730-4801-8752-e89e257c9a45?j=3DeyJ1IjoiM2dmeXZtIn0= =2Exu76uFObqArDfP822j-jnN48_jCfgM3m0rbAsF0l24U ] and business health. For marketers, it=E2=80=99s key for judging the success of their acquisitio= n efforts. For product teams, it shows the impact of product launches and changes. For engineers, it clarifies what to work on better than =E2=80=9Cgeneral us= age.=E2=80=9D For customer-facing teams, it=E2=80=99s a warning that users aren=E2=80=99t= succeeding. As cringe as it is to say it, activation creates =E2=80=9CALIGNMENT=E2=80= =9D and figuring it out can create a lot of value for your organization as = a whole. How we find activation metrics at PostHog We have 10+ products, each requiring their own activation metric, so we hav= e a lot of experience finding activation metrics [ https://substack.com/red= irect/a1ffe24f-c5c7-4895-a96d-fa366dcb946a?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFOb= qArDfP822j-jnN48_jCfgM3m0rbAsF0l24U ] that work. Our process starts by gathering a big list of events we think might "hook" = users into our product. We get these by looking at usage of that specific p= roduct, and by asking customer-facing teams what they think. Next, we create groups of 3-5 events to test together. Typically, we aim fo= r 5-10 different groups, some of which compare how often an event is comple= ted =E2=80=93 e.g. watched five replays vs. watched one replay. Once we have our potential activation metrics, we write an SQL query [ http= s://substack.com/redirect/e764e022-7278-48ab-8b24-fac04600ff1a?j=3DeyJ1Ijoi= M2dmeXZtIn0.xu76uFObqArDfP822j-jnN48_jCfgM3m0rbAsF0l24U ] to test how they = correlate with retention. The query starts by getting the companies that signed up for PostHog, broke= n down by month. We use group analytics [ https://substack.com/redirect/7b9= d4ee4-66af-4270-9e7f-d68ac658cbc4?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822= j-jnN48_jCfgM3m0rbAsF0l24U ] to find activated companies rather than indivi= dual users, but both work. Next, we filter for companies who completed our collection of activation ev= ents in the required time. The time cutoff varies by the natural product li= fecycle. It takes longer to collect enough data for product analysis than c= reate an A/B test; therefore, product analytics has a 30-day activation win= dow while experimentation only has 14 days. After finding the companies who activated successfully, we calculate their = retention by checking how many are still using the product three months aft= er starting. We then compare the activated retention against our product's = average retention rate [ https://substack.com/redirect/8fc3cabd-91a2-41ad-b= 030-ac50e5eedebe?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822j-jnN48_jCfgM3m0r= bAsF0l24U ]. In the above hypothetical example, three companies activated but only two w= ere still using the product after three months, and were therefore retained= =2E This means we have a retention rate of 66% for activated companies: If our average retention was lower than 66%, this would be a good activatio= n metric. =F0=9F=A5=B3 If our average retention was higher than 66%, this would be a bad activatio= n metric. =F0=9F=98=AD We repeat this process for each potential activation metric. The one with t= he best retention rate (and a large enough sample size) wins. Some activation metrics we found with this method: Experimentation: 1 experiment launched within 14 days. Feature flags: 2 feature flags created and 2 feature flags updated with pro= perty filters within 14 days. Product analytics: first team event ingested, 1 dashboard created, and 3 in= sight saved within 30 days. Surveys: 1 survey launched within 30 days. Session replay: 5 recordings analyzed and 1 recording list filter changed w= ithin 14 days. Want to see the full real SQL query we used for these? Check out our post o= n How we found our activation metric (and how you can too) [ https://substa= ck.com/redirect/a1ffe24f-c5c7-4895-a96d-fa366dcb946a?j=3DeyJ1IjoiM2dmeXZtIn= 0.xu76uFObqArDfP822j-jnN48_jCfgM3m0rbAsF0l24U ]. What to do after you=E2=80=99ve defined activation 1. Track it accurately Because your activation metric events now stand out, it's worth making sure= they are tracked accurately. This means: Refining when it=E2=80=99s captured. For example, is it better to capture a= n uploaded file event when the upload starts or completes? Probably complet= es because it removes errored and retried uploads. Making sure it=E2=80=99s firing properly. We found that the recording analy= zed event wasn't when we looked. This meant more teams were activating than= we realized, and our metrics reliant on activation were inaccurate.=20 Capturing it on the server. Although setting up a reverse proxy [ https://s= ubstack.com/redirect/04512cd5-bf4b-4962-900a-7e992f1c478f?j=3DeyJ1IjoiM2dme= XZtIn0.xu76uFObqArDfP822j-jnN48_jCfgM3m0rbAsF0l24U ] is a good intermediate= step, client-side events can still be ad blocked. Capturing your activatio= n events on the server [ https://substack.com/redirect/35008f45-c624-4bdc-9= 9c5-2ffa75dd879f?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822j-jnN48_jCfgM3m0r= bAsF0l24U ] ensures you get them all. 2. Review it Once we have an activation metric, it becomes part of our recurring growth = reviews [ https://substack.com/redirect/085bb9f4-ef66-4128-b874-3a2a11f358e= 3?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822j-jnN48_jCfgM3m0rbAsF0l24U ]. Fo= r every product, every month, we check: If it has dramatically changed and figure out the causes of that. Was it a = change we made? Generally, we are happy when activation stays stable. If we made changes aimed at improving activation, did they work? If not, wh= y? If yes, is there more we can do? The outcome of our growth reviews is a list of action items for the team to= work on and then review in the next meeting. 3. See how it correlates with other areas As activation is one part of the funnel [ https://substack.com/redirect/9bf= 35c10-5847-40b8-9f21-d3a619a97346?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822= j-jnN48_jCfgM3m0rbAsF0l24U ], you can use it to evaluate and improve other = parts of the funnel. For example, related to acquisition: Explore what marketing activities have an impact on activation. Target marketing and onboarding work to improve activation rates. Improve sales by using (lack of) activation as a sales lead health check. And related to retention, revenue, and referral: Look at activation's impact on retention beyond three months. Using activation to look for opportunities to do sales-assist. Finding ways to convert activated users into referrers. A special, related one for us (and other multi-product companies) is cross-= sell. Getting people to activate into multiple products (like product analy= tics + feature flags) is extremely valuable to us, because they are much mo= re likely to pay, and pay more when they do. Because of the importance of cross-sell, we look at how the activation of o= ne product impacts another, figure out the "right" products to promote for = a user's lifecycle, and look for opportunities to promote "cross-activation= =2E"=20 This is a big area of focus for our growth team. 4. Just improve it Eventually, you need to actually build something to help activate users fas= ter or provide them more motivation to activate. Some examples: Creating email campaigns [ https://substack.com/redirect/c3e467de-2c04-41b8= -a337-1c196bd2e66d?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822j-jnN48_jCfgM3m= 0rbAsF0l24U ] focused on your activation events Tracking signup to activation conversion with funnels [ https://substack.co= m/redirect/e06eb25b-8cd8-4185-91c9-9e5a8f76d3b7?j=3DeyJ1IjoiM2dmeXZtIn0.xu7= 6uFObqArDfP822j-jnN48_jCfgM3m0rbAsF0l24U ] Running experiments [ https://substack.com/redirect/79011c52-534d-451f-8123= -a147cb49e43a?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822j-jnN48_jCfgM3m0rbAs= F0l24U ] with your activation events as a goal metric Launching surveys [ https://substack.com/redirect/9fffacfb-5090-4997-9b5f-f= f2ff875b84a?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822j-jnN48_jCfgM3m0rbAsF0= l24U ] asking people why they don't activate Connecting activation to sales or revenue data with our data warehouse [ ht= tps://substack.com/redirect/e3b73c97-bee3-48bd-aaaf-e75a1c451af6?j=3DeyJ1Ij= oiM2dmeXZtIn0.xu76uFObqArDfP822j-jnN48_jCfgM3m0rbAsF0l24U ] In one case for us, because the =E2=80=9Creplay list filter added=E2=80=9D = event is part of our activation metric, we focused on changes that encourag= e users to use playlists like our "What to watch" tab with playlist templat= es: Paul Copplestone [ https://substack.com/redirect/587e2060-a709-435a-9bf7-ec= 3a05e53071?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822j-jnN48_jCfgM3m0rbAsF0l= 24U ], CEO of Supabase [ https://substack.com/redirect/9262d13a-3f50-4d66-a= 4e8-dd19eb1aed3e?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822j-jnN48_jCfgM3m0r= bAsF0l24U ], gave me a great answer when I asked how they tried to improve = their activation metric of "weekly active databases": Is it ever 1 thing? This metric has many drivers: increase new databases (b= y increasing new signups or make it compelling for existing signups to spin= up many databases), reduce churn by ensuring no scalability issues, increa= se the growth of the database from "just trying it out" to "using in produc= tion" by reducing friction when building, etc. Activation, for everyone, is a metric that is both business-critical and in= finitely optimizable. For teams not sure how to use analytics to improve th= eir product, this makes it an excellent place to start. Words by Ian Vanagas, who is very serious about activation metrics. Thanks for reading! Help me optimize this email=E2=80=99s activation rate b= y sharing with a friend =F0=9F=A7=A0 Good reads to further activate your mind The half-life of code & the ship of Theseus [ https://substack.com/redirect= /3eb4ac26-2313-42dc-be59-9e229d52e369?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFObqArDf= P822j-jnN48_jCfgM3m0rbAsF0l24U ] =E2=80=93 Erik Bernhardsson =E2=80=9CStealth mode=E2=80=9D and other f=E2=80=99ing brilliant strategies= [ https://substack.com/redirect/ae22da30-1b6c-4d15-a8b5-e25dd35c026f?j=3De= yJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822j-jnN48_jCfgM3m0rbAsF0l24U ] =E2=80=93 = Jason Cohen Confessions of a Big Tech recruiter [ https://substack.com/redirect/da34e65= c-87ab-44a4-9ab8-a8f8022c7f28?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822j-jn= N48_jCfgM3m0rbAsF0l24U ] =E2=80=93 Gergely Orosz =20 How to launch a dev tool on Hacker News [ https://substack.com/redirect/845= d38a8-c1ca-4581-b45b-9388232a6c89?j=3DeyJ1IjoiM2dmeXZtIn0.xu76uFObqArDfP822= j-jnN48_jCfgM3m0rbAsF0l24U ] =E2=80=93 Jakub Czakon Unsubscribe https://substack.com/redirect/2/eyJlIjoiaHR0cHM6Ly9uZXdzbGV0dGV= yLnBvc3Rob2cuY29tL2FjdGlvbi9kaXNhYmxlX2VtYWlsP3Rva2VuPWV5SjFjMlZ5WDJsa0lqb3= lNRGt3TVRjME1qWXNJbkJ2YzNSZmFXUWlPakUxTmpNNU9EZ3pNU3dpYVdGMElqb3hOek00T0RZM= k5EVXlMQ0psZUhBaU9qRTNOekEwTURJME5USXNJbWx6Y3lJNkluQjFZaTB4TXpFNE1qSTFJaXdp= YzNWaUlqb2laR2x6WVdKc1pWOWxiV0ZwYkNKOS5QblREcF96V2V5cmZ5ZU1icVJiNUVGSG5BWEt= EY0tNekdJb0hvdW5zeENVIiwicCI6MTU2Mzk4ODMxLCJzIjoxMzE4MjI1LCJmIjp0cnVlLCJ1Ij= oyMDkwMTc0MjYsImlhdCI6MTczODg2NjQ1MiwiZXhwIjoxNzQxNDU4NDUyLCJpc3MiOiJwdWItM= CIsInN1YiI6ImxpbmstcmVkaXJlY3QifQ.tTuwCn_Ckuk05g_eNykS-LB2MXgIwCrjOGAq8VYlO= -E? --2e42b96d8f1b7cd70d36317e2fc25470063070b987f96e53df900dad4967 Content-Type: text/html; charset="utf-8" Content-Transfer-Encoding: quoted-printable WTF is activation and why should = engineers care?
An engineer's guide to find= ing (and using) your product's activation metric
͏     = 73;͏     ­͏     ­͏   &#= 8199; ­͏     ­͏     ­͏ =     ­͏     ­͏     = 3;͏     ­͏     ­͏   = 199; ­͏     ­͏     ­͏ &= nbsp;   ­͏     ­͏     ­= ;͏     ­͏     ­͏   Q= 99; ­͏     ­͏     ­͏ &n= bsp;   ­͏     ­͏     ­= ͏     ­͏     ­͏   ̳= 9; ­͏     ­͏     ­͏ &nb= sp;   ­͏     ­͏     ­&= #847;     ­͏     ­͏    = ; ­͏     ­͏     ­͏ &nbs= p;   ­͏     ­͏     ­&#= 847;     ­͏     ­͏    = ­͏     ­͏     ­͏  = ;   ­͏     ­͏     ­= 47;     ­͏     ­͏     = ­͏     ­͏     ­͏  =   ­͏     ­͏     ­T= 7;     ­͏     ­͏     &= #173;͏     ­͏     ­͏   =   ­͏     ­͏     ­͏= ;     ­͏     ­͏     &#= 173;͏     ­͏     ­͏   &= #8199; ­͏     ­͏     ­͏=     ­͏     ­͏     = 73;͏     ­͏     ­͏   &#= 8199; ­͏     ­͏     ­͏ =     ­͏     ­͏     = 3;͏     ­͏     ­͏   = 199; ­͏     ­͏     ­͏ &= nbsp;   ­͏     ­͏     ­= ;͏     ­͏     ­͏   Q= 99; ­͏     ­͏     ­͏ &n= bsp;   ­͏     ­͏     ­= ͏     ­͏     ­͏   ̳= 9; ­͏     ­͏     ­͏ &nb= sp;   ­͏     ­͏     ­&= #847;     ­͏     ­͏    = ; ­͏     ­͏     ­͏ &nbs= p;   ­͏     ­͏     ­&#= 847;     ­͏     ­͏    = ­͏     ­͏     ­͏  = ;   ­͏     ­͏     ­= 47;     ­͏     ­͏     = ­͏     ­͏     ­͏  =   ­͏     ­͏     ­T= 7;     ­͏     ­͏     &= #173;͏     ­͏     ­͏   =   ­͏     ­͏     ­͏= ;     ­͏     ­͏     &#= 173;͏     ­͏     ­͏   &= #8199; ­͏     ­͏     ­͏=     ­͏     ­͏     = 73;͏     ­͏     ­͏   &#= 8199; ­͏     ­͏     ­͏ =     ­͏     ­͏     = 3;͏     ­͏     ­͏   = 199; ­͏     ­͏     ­͏ &= nbsp;   ­͏     ­͏     ­= ;͏     ­͏     ­͏   Q= 99; ­͏     ­͏     ­͏ &n= bsp;   ­͏     ­͏     ­= ͏     ­͏     ­͏   ̳= 9; ­͏     ­͏     ­͏ &nb= sp;   ­͏     ­͏     ­&= #847;     ­͏     ­͏    = ; ­͏     ­͏     ­͏ &nbs= p;   ­͏     ­͏     ­&#= 847;     ­͏     ­͏    = ­͏     ­͏     ­͏  = ;   ­͏     ­͏     ­= 47;     ­
Forwarded t= his email? Subscribe here for more

WTF is activation and why should e= ngineers care?

An engineer's guide to finding (and using) your product's activati= on metric

<= td align=3D"right">
=
<= div class=3D"pencraft pc-reset color-primary-zABazT line-height-20-t4M0El f= ont-meta-MWBumP size-11-NuY2Zx weight-medium-fw81nC transform-uppercase-yKD= gcq reset-IxiVJZ meta-EgzBVA custom-css-email-post-author" style=3D"list-st= yle: none;font-size: 11px;line-height: 20px;text-decoration: unset;color: r= gb(54,55,55);margin: 0;font-family: 'SF Compact',-apple-system,system-ui,-a= pple-system,BlinkMacSystemFont,'Segoe UI',Roboto,Helvetica,Arial,sans-serif= ,'Apple Color Emoji','Segoe UI Emoji','Segoe UI Symbol';font-weight: 500;te= xt-transform: uppercase;letter-spacing: .2px;">Ian Vanagas
<= /table>
 
<= tbody>=
3D""
3D""
3D""
=3D""
READ IN APP3D""
 

Welcome = to Product for Engineers!

PostHog recently moved you from another pla= tform to Product for Engineers, hosted on Substack. New posts will be auto= matically delivered to you via email or via the Substack app.
To se= t up your profile and find more to read, click here. If you do not want to= receive emails from Product for Engineers, you can unsubscribe here.

Improving activation is one of the most impactful = things a produ= ct engineer can do, but it’s also weirdly hard to define.

People often call it your product’s "aha" momen= t – i.e. when a user experiences your product's value for the first t= ime – but this is vague and subjective.

So why i= sn't there a simple, universal definition? Because:

  1. It’s unique to you= r product. No two products are exactly the same, so it’= ;s impossible to generalize. This is also why activation benchmarks suck.¹

  2. It’s rarely a single event. You= want users to do many actions in your product: complete onboarding, upload= a file, share a link, watch a video. Activation could mean doing any of th= ese, or a combination of them.

  3. Activating may mean doing something multiple times. Someon= e might need to do the action multiple times in order to get it. For exampl= e, to activate into our session replay product means analyzing not <= em>one, but five replays.

    4. <= /ol>

    One thing is certain, though. You can’t build a succ= essful product without nailing it. Read on to understand why.

    Why engineers sho= uld care about activation

    1. It's upstream of other metrics

    In a way, activation is the base rate of your product.

    Want more revenue? Improve activation. Weak word-of-mouth growth getti= ng you down? Improve activation. It compounds all the metrics downstream of= it.

    If being a product engineer means caring about t= he whole business (spoiler: it does), then you should be obsessed with your pro= duct’s activation metric.

    A larger sample size= (i.e. more activated users) also makes it easier to analyze and experiment= on downstream metrics.

    TL;DR: Every= thing is easier when activation is strong.

    2. It's something eng= ineers can directly influence

    =
    3D""

    Activation is part of the AARRR funnel. Above it, at the top of the funnel, = is acquisition which is mostly a responsibility of marketers.

    Below it is retention, revenue, and referral. The last two are mo= stly platform-level metrics, and normally the responsibility of a smaller g= rowth team.

    That leaves activation and retention as the responsibili= ties of product engineers. Activation comes first and directly influences r= etention, so it’s a natural starting point.

    3. It’s = useful across the organization

      =

    • For founders, it= 217;s a signal of product market-fit and business health.

    • For marketers, it’s key f= or judging the success of their acquisition efforts.

    • For product teams, it shows the impact = of product launches and changes.

    • For engineers, it clarifies what to work on better than = 220;general usage.”

    • For customer-facing teams, it’s a warning that users aren= 217;t succeeding.

    As cringe as it is to say it, a= ctivation creates “ALIGNMENT” and figuring it out can create a = lot of value for your organization as a whole.

    How we find activation metrics a= t PostHog

    3D""

    We have 10+ = products, each requiring their own activation metric, so we have a lot of e= xperience find= ing activation metrics that work.

    Our proc= ess starts by gathering a big list of events we think might "hook" users in= to our product. We get these by looking at usage of that specific product, = and by asking customer-facing teams what they think.

    Next, = we create groups of 3-5 events to test together. Typically, we aim for 5-10= different groups, some of which compare how often an event is completed &#= 8211; e.g. watched five replays vs. watched one replay.

    = 3D""
    =

    Once we have our potential activation met= rics, we write an SQL query to test how they correlate with retention.

    The query starts by getting the companies that signed = up for PostHog, broken down by month. We use group analytics to find activated = companies rather than individual users, but both work= =2E

    Next, we filter for companies who completed our co= llection of activation events in the required time. The time cutoff varies = by the natural product lifecycle. It takes longer to collect enough data fo= r product analysis than create an A/B test; therefore, product analytics ha= s a 30-day activation window while experimentation only has 14 days.

    After finding the companies who activated successfully, we = calculate their retention by checking how many are still using the product = three months after starting. We then compare the activated retention agains= t our product's

    3D"Activation"

In the above hypothetical exam= ple, three companies activated but only two were still using the product af= ter three months, and were therefore retained. This means we have a retenti= on rate of 66% for activated companies:

We repeat this process for each potential a= ctivation metric. The one with the best retention rate (and a large enough = sample size) wins.

Some activation metrics we found with th= is method:

=

Want to see the full real SQL query = we used for these? Check out our post on How we found our activation metr= ic (and how you can too).

What to do after you= 217;ve defined activation

3D"Dog=
<= /div>

1. Track it accurately

Because your activation metri= c events now stand out, it's worth making sure they are tracked accurately.= This means:

  1. Refining when it’s captured. For example, = is it better to capture an uploaded file event wh= en the upload starts or completes? Probably completes because it removes errored and= retried uploads.

  2. Making sure= it’s firing properly. We found that the = recording analyzed event wasn't when we looked. This meant mor= e teams were activating than we realized, and our metrics reliant on activa= tion were inaccurate.

  3. <= p style=3D"color: rgb(54,55,55);line-height: 26px;margin-bottom: 0;box-sizi= ng: border-box;padding-left: 4px;font-size: 16px;margin: 0;">Captur= ing it on the server. Although setting up a reverse proxy is a g= ood intermediate step, client-side events can still be ad blocked. Capturin= g your activation events on the server ensures you get them all.

2. Review it

Once we have an activation metri= c, it becomes part of our recurring growth reviews. For every product, every mo= nth, we check:

  • If it has dramatically changed and= figure out the causes of that. Was it a change we made? Generally, we are = happy when activation stays stable.

  • If we made changes aimed at improving activation, did th= ey work? If not, why? If yes, is there more we can do?

The outcome of our growth reviews is a list of action items for the t= eam to work on and then review in the next meeting.

3. See how it corre= lates with other areas

As activation is one part of = the funnel= , you can use it to evaluate and improve other parts of the funnel. F= or example, related to acquisition:

  • <= p style=3D"color: rgb(54,55,55);line-height: 26px;margin-bottom: 0;box-sizi= ng: border-box;padding-left: 4px;font-size: 16px;margin: 0;">Explore = what marketing activities have an impact on activation.²

  • Target marketing and onboarding work to improve a= ctivation rates.

  • = Improve sales by using (lack of) activation as a sales lead health check.

And related to retention, revenue, and referral:

  • Look at activation's impact on retention beyond three mont= hs.

  • =

    Using activat= ion to look for opportunities to do sales-assist.

  • Finding ways to convert activated users in= to referrers.

A special, related one for us= (and other multi-product companies) is cross-sell<= span>. Getting people to activate into multiple products (like product anal= ytics + feature flags) is extremely valuable to us, because they are much m= ore likely to pay, and pay more when they do.

Becaus= e of the importance of cross-sell, we look at how the activation of one pro= duct impacts another, figure out the "right" products to promote for a user= 's lifecycle, and look for opportunities to promote "cross-activation."

This is a big area of focus for our growth team.

4. Just = improve it

3D"Improve"

Eventual= ly, you need to actually build something to help activate users faster or p= rovide them more motivation to activate. Some examples:

In one case for us,= because the “replay list filter added” event is part of our ac= tivation metric, we focused on changes that encourage users to use playlist= s like our "What to watch" tab with playlist templates:

= 3D"What

Paul C= opplestone, CEO of Supabase, gave me a great answer when I asked how = they tried to improve their activation metric of "weekly active databases":=

Is it ever 1 thing? This met= ric has many drivers: increase new databases (by increasing new signups or = make it compelling for existing signups to spin up many databases), reduce = churn by ensuring no scalability issues, increase the growth of the databas= e from "just trying it out" to "using in production" by reducing friction w= hen building, etc.

Activation, for everyone, i= s a metric that is both business-critical and infinitely optimizable. For t= eams not sure how to use analytics to improve their product, this makes it = an excellent place to start.

Words by Ian Vanagas, who = is very serious about activation metrics.

Thanks for rea= ding! Help me optimize this email’s activation rate by sharing with a= friend

<= a class=3D"button primary" rel=3D"" href=3D"https://substack.com/app-link/p= ost?publication_id=3D1318225&post_id=3D156398831&utm_source=3Dsubstack&utm_= medium=3Demail&utm_content=3Dshare&utm_campaign=3Demail-share&action=3Dshar= e&triggerShare=3Dtrue&isFreemail=3Dtrue&r=3D3gfyvm&token=3DeyJ1c2VyX2lkIjoy= MDkwMTc0MjYsInBvc3RfaWQiOjE1NjM5ODgzMSwiaWF0IjoxNzM4ODY2NDUyLCJleHAiOjE3NDE= 0NTg0NTIsImlzcyI6InB1Yi0xMzE4MjI1Iiwic3ViIjoicG9zdC1yZWFjdGlvbiJ9.Gqod2_h85= l_s0PWjuDoGB3iH7Ubj7Ctw_0Vrh0urXO8" style=3D"font-family: system-ui,-apple-= system,BlinkMacSystemFont,'Segoe UI',Roboto,Helvetica,Arial,sans-serif,'App= le Color Emoji','Segoe UI Emoji','Segoe UI Symbol';display: inline-block;bo= x-sizing: border-box;cursor: pointer;border: none;border-radius: 8px;font-s= ize: 14px;line-height: 20px;font-weight: 600;text-align: center;margin: 0;o= pacity: 1;outline: none;white-space: nowrap;color: #ffffff !important;text-= decoration: none !important;background-color: #F54E00;padding: 12px 20px;he= ight: auto;">Share

🧠 Good reads to further activate your mind

<= /td>
1

What about my benchmarks bro? Lenny's activation post also has some nice benchmarks for activation, b= ut the ranges are extremely wide. Because the definition of "activation" va= ries between businesses, there's no real way to compare. When you are battl= ing to increase your activation rates by single digit percentage points, ha= ving a range of 20%+ between "ok" and "good" isn't helpful.

<= span class=3D"footnote-number" style=3D"display: block;margin-right: 6px;mi= n-width: 24px;">2

We fo= und that viewing two web pages (e.g. our docs and product pages) before sig= ning up led to a dramatically higher activation rate, compared to users who= went straight from our homepage, or pricing page, to signing up.

 
<= td width=3D"8" style=3D"min-width: 8px">
= <= /tr>
3D""Like
3D""Comment
3D""Restack
<= tbody>
 

© 2025 PostHog
548 Market St= reet PMB 72296, San Francisco, CA 94104
Unsubscribe

3D"Get3D"Start

3D""3D"" --2e42b96d8f1b7cd70d36317e2fc25470063070b987f96e53df900dad4967--